20100323Kopzcynski Di.pdf

DAVID J. MEYER
VICE PRESIDENT AN CHIEF COUNSEL OF
REGULATORY & GOVERNTAL AFFAIRS
AVISTA CORPORATION
P.O. BOX 3727
1411 EAST MISSION AVENUE
SPOKA, WASHINGTON 99220 - 3 727
TELEPHONE: [phone]
FACSIMILE: [phone]
DAVID. MEYER~AVISTACORP. COM
') ". '"t
,\ c~ J f!: OS
BEFORE TH IDAO PUBLIC UTILITIES COMMISSION
IN THE MATTER OF THE APPLICATION )
OF AVISTA CORPORATION FOR THE )
AUTHORITY TO INCREASE ITS RATES )
AN CHAGES FOR ELECTRIC AN )
NATUR GAS SERVICE TO ELECTRIC )
AN NATURA GAS CUSTOMERS IN THE )STATE OF IDAHO )
)
CASE NO. AVU-E-10-01
CASE NO. AVU-G-10-01
DIRECT TESIMONY
OF
DON F. KOPZCYNSKI
FOR AVISTA CORPORATION
(ELECTRIC AN GAS)
1
2 I.INTODUCTION
3 Q.Please state your nae, emloyer an business
4 address.
5 A.My name is Don F. KopczYnski and I am employed as
6 the Vice President of Transmission and Distribution Operations
7 for Avista Utilities, at 1411 East Mission Avenue, Spokane,
8 washington.
9 Q.Would yo briefly describe your educational
10 background aDd professional exerience?
11 A.Yes. Prior to joining the Company in 1979, I earned
12 a Bachelor of Science Degree in Engineering from the
13 Uni versi ty of Idaho.I have also earned a Mas ter ' s Degree in
14 Engineering from Washington State University and a Master's
15 Degree in Organizational Leadership from Gonzaga University.
16 Over the past 31 years I have spent approximately 17 years in
17 Energy Delivery, managing Engineering, various aspects of
18 Operations, and Customer Service.In addition, I spent three
19 years managing the Energy Resources Department, including
20 Power Supply, Generation and Production, and Natural Gas
21 Supply.I have worked in the areas of Corporate business
22 analysis and development,and served in a variety of
23 leadership roles in subsidiary operations for Avista Corp.I
KopczYnski, Di
Avista Corp
Page 1
1 was appointed General Manager of Energy Delivery in 2003 and
2 vice President in 2004.I serve on several boards, including
3 the washington State Electrical Board,Northwest Gas
4 Association, American Gas Association, Common Ground Alliance
5 and the Washington State University Engineering Advisory
6 Board.
7 Q.Wht is the scope of your testimny?
8 A.i will provide an overview of the Company's electric
9 and natural gas energy delivery facilities and operations. I
10 will also explain some of our efforts to control costs,
11 increase efficiency, and improve customer service, as well as
12 sumarize Avista's customer support programs in Idaho.A
13 table of the contents for my testimony is as follows:
14 Description Page
15
16
17
18
19
i.II.III.
iv.
Introduction 1
Overview of Avista' s Energy Delivery Service 3
Cost Control and Efficiency Efforts 7Customer Support Programs 13
20 Q.Are you sponsoring any exibits in this proceeding?
21 A.Yes.I am sponsoring Exliibi t No. 7 Schedule 1 and
22 Schedule 2. Schedule 1 shows the detailed usage and numer of
23 customers for each customer class. Schedule 2 is a 2009 study
24 performed at Eastern Washington University addressing heating
KopczYnski, Di
Avista Corp
Page 2
1 assistance programs in our service area. These exibits were
2 prepared under my direction.
3
4 II. OVEVJEW OF AVJSTA'S ENRGY DELIVERY SERVJCE
5 Q. Please describe Avista utilities' idah electric and
6 natural gas utility operations.
7 A.Avista Utilities operates a vertically-integrated
8 electric system. In addition to the hydroelectric and thermal
9 generating resources described by Company witness Mr. Storro,
10 the Company has approximately 4,052 miles of lines in the
11 following classes in Idaho: 286 miles of 230 kV transmission,
12 604 miles of 115 kV transmission, and 3,162 miles of sub-
13 transmission and distribution line at a variety of voltages.
14 Avista also has 928 miles of distribution underground cable;
15 the predominant distribution voltage is 13.2 kV. Avista owns
16 and maintains 1876 miles of natural gas pipelines (excluding
17 services) in the state of Idaho of which 560 miles are steel
18 and 1316 miles are polyethylene.All of these pipelines are
19 distribution, not transmission, operating at various maximum
20 allowable operating pressures (MAOPs) from 60 psig to 720
21 psig.Avista has 69,337 natural gas service lines in Idaho.
22 A map showing the Company's electric and natural gas service
KopczYnski, Di
Avista Corp
Page 3
1 area in Idaho is provided by Mr. Morris at page 2 of Exhibit
2 No. 1.
3 As detailed in the Company's 2009 Electric integrated
4 Resource Plan, Avista expects retail electric sales growth to
5 average 1.7% .annually for the next ten years and 1.7% over the
6 next twenty years in Avista's service territory, primarily due
7 to increased population and business growth. In 2008, Avista
8 had 4,493 new electric customer connections1 and 3,350 for
9 2009. A copy of the Company's 2009 Electric IRP has been
10 attached as Exhibi t NO. 4 Schedule 1 to Mr. Storro' s
11 testimony.
12 Also, based on Avista' s 2009 Natural Gas Integrated
13 Resource Plan, in Idaho/Washington the numer of customers
14 were projected to increase at an average annual rate of 2.2%,
15 with demand growing at a compounded average annual rate of
16 1.0%. New natural gas customer connections were 4,797 in 2008
17 and 3,362 in 2009. A copy of the Company's 2009 Natural Gas
18 IRP has been attached as Exhibi t No. 11 , Schedul e 2 to Mr.
19 Christie's testimony.
1 A new customer connection as defined by Avista is when a customer
receives a bill for the first time at a particular premise/location.
KopczYnski, Di
Avista Corp
Page 4
1 Q.Bow may customers are served bY Avista utilities in
2 Idaho?
3 A.Of the Company's 356,620 electric and 316,350
4 natural gas customers (as of Decemer 31, 2009), 122,358 and
5 74,006, respectively, were Idaho customers.Avista's largest
6 electric customer in Idaho is Clearwater Paper, located in
7 Lewiston, Idaho.
8 Q.please describe the Comany's operations centers
9 that support electric an natural gas customrs in Idah.
10 A.The Company has construction offices in Grangeville,
11 Orofino,Lewiston-Clarkston,Moscow-pullman,Kellogg,St.
12 Maries, Coeur d' Alene, Sandpoint and Bonner's Ferry, and
13 customer contact center operations in Lewiston and Coeur
14 d' Alene. Avista' s four customer contact centers in Coeur
15 d' Alene, Lewiston, Spokane, and Medford, Oregon are networked,
16 allowing the full pool of regular and part-time employees to
17 respond to customer calls in all jurisdictions.
18 Q.Wht construction an maintenance programs does the
19 Comany have in place to maintain natural gas aD electric
20 facilities?
21 A.The Company utilizes seasonal and regular crews for
22 natural gas and electric construction, including new and
23 reconstructed lines,damage repair,and connecting new
KopczYnski, Di
Avista Corp
Page 5
1 customers.The Company employs contract crews and temporary
2 and part-time employees to meet customer needs during the peak
3 construction season. The Company also has several maintenance
4 programs to maintain the reliability of our electric and
5 natural gas infrastructure.On the electric side, this
6 includes the Company's asset management program (including
7 wood pole inspection and replacement), vegetation management,
8 electric transmission line inspection and reconstruction.
9 Company witness Mr. Kinney discusses this program in more
10 detail.
11 Regarding natural gas operations, ongoing maintenance
12 focuses on valve and regulator stations, atmospheric and
13 underground corrosion protection, and leak surveys.. Natural
14 gas operations performs necessary maintenance required by the
15 US Department of Transportation Pipeline Safety Regulations,
16 49 CFR,Part 192.Emergency valves are inspected and
17 maintained to make sure they are accessible for operation,
18 they turn satisfactorily,and are identified properly.
19 Atmospheric Corrosion Inspection is performed on all of our
20 above-ground piping facilities at least every three years. To
21 levelize the workload, approximately one third of our system
22 is maintained annually.Piping is inspected to assure it is
23 coated properly to protect against corrosion.underground
KopczYnski, Di
Avista Corp
Page 6
1 corrosion protection surveys are performed annually on
2 underground steel piping. Rectifiers that induce current onto
3 the pipe to supply corrosion protection are inspected six
4 times per year.Additionally, whenever a buried steel
5 pipeline is exposed, crews inspect the pipe for coating
6 deterioration and external corrosion.
7 Finally,leak Surveys are performed at differing
8 intervals,with facilities in more populated "business
9 districts" inspected annually, and those in less populated
10 residential areas are inspected every five years.
11
12 III. COST CONTROL AN EFFICIENY EFFORTS
13
14 Q.Given the current aD near-ter. economc coDdi tions,
15 what actions or specific measures has the Comany unertaken
16 to control costs aD mitigate the requested rate increase?
17 A.As Mr. Morris noted in his testimony, following the
18 energy crisis of 2000/2001, we cut our operating expenses as
19 we worked toward regaining an investment grade credit rating.
20 Since that time we have continued to pay particular attention
21 to limiting the growth in these costs, while meeting important
22 reliability and environmental compliance requirements, and
23 preserving a high level of customer satisfaction.
KopczYnski, Di
Avista Corp
page 7
1 The measures listed below are among some of the most
2 recent actions we have taken to mitigate the impact of
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
increased costs on our customers:
1. Limitations on Capital Spening. For both 2009 and 2010
Avista approved a lower capital budget than was
requested by the Company's Engineering and Operationspersonnel. The Capital Prioritization Co~ittee
reduced the list of projects to be completed by
approximately $60 million in 2009, and we have limited
our near-term capital budget to approximately $210
million annually.
2. Biring .Restriction. The Company continues to operate
under a hiring restriction which requires approval by
the Chairman/President/CEO, CFO, and Sr. VP for Human
Resources for all replacement or new hire positions.
3. improvements and Efficiency initiatives. Avista
Utilities has undertaken a numer of improvements and
efficiency initiatives throughout our service area that
are focused on either increasing customer service and
satisfaction, or increasing productivity and reducing
operating costs. These measures have served to
mitigate the impact on customers of the proposed rate
increase. In regard to Information Services, Avista
has been, and continues to be, focused on cost
effective solutions that meet our customers' needs.
One way to meet a growing customer demand for
transaction choices is through the appropriate use of
technology. Most recently, Avista has focused on
reducing customer transaction costs through the use of
technology, such as the Outage Management Tool (OMT)
which enables a customer to report outages without
talking to a representative.
Q. Did the Comauy initiate a numr of cost magemnt
initiatives in the years just prior to the recent dowturn in
the econom?
KopczYnski, Di
Avista Corp
Page 8
1 A.Yes. Avista's efforts to control its costs have not
2 been prompted solely by the most recent downturn in the
3 economy.We have continually revisited our costs and
4 operating practices over time in order to mitigate price
5 increases for our customers.A sampling of other measures
6 that we had already taken prior to the downturn in the economy
7 include the following:
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
.Retirees are now picking up the full premium
increases on the health insurance coverage. A few
years ago retirees under age 65 were paying 10% of
the health insurance premiums and now they pay 50%
on average.
.The Defined Benefit Pension Plan's benefit
formulas were reduced (approximately 28%) for all
new hires effective January 1, 2006 and forward.
This applies to all new hires except those in the
IBEW Local #77 Bargaining Unit.
.Bargaining units wages were kept in line with
neighboring investor-owned utilities and PUDs.
.Normally Avista will bring on about 15 to 18
temporary groundsmen in the Spring to assist in
the construction work for the remainder of the
year. This pool of people helps us manage through
the construction season with new developments that
take place from April to December. We use this
pool of people to select upcoming line apprentices
in anticipation of future retirements. Due to the
downturn in the economy and the lack of new
construction projects, the Company decided to not
hire temporary groundsmen for the year unless a
specific project would dictate a need for one or
two people. The savings for 2009 was
approximately $700,000. Assuming a more normal
construction season in 2010, we will be back to
KopcZYnski, Di
Avista Corp
Page 9
1
2
3
4
5
6
7
8
9
10
11
12
.13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
our normal practice and bring on the groundsmen
for the construction season. Therefore, the
savings to the Company is short term versus
annual on-going savings.
. Additionally, regarding our natural gas
operations, approximately 10% of our natural gasconstruction workforces are classified' as
"seasonal. " Seasonal and temporary employees are
let go at the end of the construction season and
brought back in March or April as construction
starts to ramp up. In 2009, we delayed bringing
back these employees due to the downturn in
construction and won' t re-employ these workers
unless construction activity improves. Again,
these saving occurred in 2009 and are not
anticipated to carry over into 2010 unless the
economy and construction continues to be slow.
. Starting in 2007 the Company has also realized
further efficiencies in employee training:
o Shortened the natural gas apprenticeship time by
12 to 18 months by bringing in advance standing
employees who already have the skills andabilities;
o Reduced the anual natural gas refresher trainingrequired by PHMSA for Operator Qualification
through the use of on-line training programs by
one full day, and eliminated additional instructor
travel time and expense during the remainder of
the year. Estimated savings are approx 150 gas
employees at 8 hours. The trainer savings for the
remainder of the year is 20+ travel days for each
of the instructors;
o Combined different apprentice training programs in
the generation and electric areas to save over 100
hours of instructor time;
o Utilized retired craft employees for pre-
apprentice line school and other apprentice
program training, saving benefit costs and
utilizing flexible hours;
o Provided an on-site physical therapist to shorten
medical treatment time for employees as well as
reduce time away from work for medical
KopczYnski, Di
Avista Corp
Page 10
1
2
3
4
5
6
7
8
9
10
11
appointments.
. The Company has increased shift coverage company-
wide for natural gas and electric servicemen for
after (normal) hour's calls. This provides for
more prompt call response at lower cost (straight
time versus overtime) .
These programs are just examples of the extensive efforts
12 of Avista to identify and implement efficiency measures and/or
13 productivity across the organization, while continuing to
14 provide quality service to customers.
15 Avista also has a numer of ongoing process improvement
16 measures related to customer service that have provided
17 savings and efficiencies as described below.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
. Avista' s Customer Service Analyst Team constantly
challenges themselves to find ways to improve the
business without compromising customer
satisfaction. Initiatives such as automating
address returns with the. US postal Service,
reviewing collection notice parameters,
implementing email management processes, improving
system response time, designing a comprehensive
screen view, ebill promotions and other
miscellaneous improvements resulted in over $1
million of productivity savings from 2004-2009.
Examples included within the $1 million in savings
include options that give customer more choices
such as:
o E-bill - 66,582 customers enrolled - Savings $.50
per bill per month.
o Web payment process - reduced company cost from
$.80 to $.10 per transaction - 50,000 transactions
per month.
KopczYnski, Di
Avista Corp
Page 11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
. In mid-2009, Avista implemented its new Enterprise
Voice Portal (EVP) System. The new EVP systemreplaced the Company's old Integrated Voice
Response (IVR) system, installed in 1997, which
was no longer being supported by the vendor. Theold IVR and new EVP systems handled 735,000
.customer calls in 2009 (approximate offset of 38
FTE' s) . This is 43.3% of the total inbound calls
into Avista. For the first two months in 2010,the EVP system handled 124,682 calls; this
represents 47.3% of inbound calls to Avista. The
new EVP system has several new features that will
increase customer self service capabilities and
improve customer satisfaction.
The following table shows that the avoided labor
18 savings from the IVR/EVP system from 1998 through 2009 totals
19 $17.5 million:
20
21
22
23
24
25
26
27
28
29
30
KopczYnski, Di
Avista Corp
Page 12
Added Account Recap self-serce270,416$
$
$
$
$
$
$
$
5.184,889
158,353
214,828
294,609
1998
504,4379.61999
684,33913.02000
Added Payment Arangement self-
serce938,48317.82001
1,093,01620.7343,1202002
1,411,807 Added Electronic Payment self-serce26.7443,1952003
1,280,80524.3402,0712004
Enhance Payment Arangement self-
1,854,079 serce22.0530,7482005
$ 2,098,550
$ 2,182,715
34.2600,7302006
30.5624,8232007
$ 2,348,82236.2682,7972008
$ 2,880,167 New EVP 1m lementation June, 200938.9735,9382009
1
2
iv. CUSTO SUPPORT PROGRAS3
Please explain the customer support programs that4Q.
5 Avista provides for its customers in Idaho.
Avista Utilities offers a numer of programs for6A.
7 its Idaho customers, such as energy efficiency programs,
8 Proj ect Share for emergency assistance to customers, a
9 Customer Assistance Referral and Evaluation Service (CARES)
KopczYnski, Di
Avista Corp
page 13
1 program, level pay plans, and payment arrangements.Some of
2 these programs will serve to mitigate the impact on customers
3 of the proposed rate increase.
4 Avista Utilities actively participated in the energy
5 affordability workshops in Case No. GNR-U-08-01. In that case,
6 workshop participants explored ways to address energy
7 affordability and the ability of customers to pay energy
8 bills. The Company worked with Staff and other interested
9 stakeholders to support legislation in the previous session
10 that would -have allowed the Commission to adopt programs such
11 as the Company's Low Income Rate Assistance Program (LIRAP) as
12 is currently in place in Washington and Oregon. That measure
13 failed, however after many weeks of various meetings, and
14 achieving agreement on numerous amendments to the original
15 bill, a revised bill was nit reintroduced in the current
16 session.
17 However, the Company and other stakeholders hope to
18 possibly introduce a revised bill in the next session.
19 Avista is also actively involved in supporting community
20 human services programs that provide tools and resources for
21 individuals and families who face challenges in meeting the
22 basic costs of living, which includes the cost of energy.
23 Through philanthropic contributions and employee community
KopczYnski, Di
Avista Corp
Page 14
1 outreach efforts, we support programs that address basic
2 needs.Avista is a strong supporter of united Way in Idaho,
3 providing corporate and employee support for the humn
4 services agencies in our service territories.
5 Q.Bas the Comau done any recent research with regard
6 to seniors an limted incom customrs it serves?
7 A. Yes. Avista, along with the low income and senior
8 advocates, has long sought to understand the reach and
9 effectiveness of energy assistance and energy efficiency
10 programs. The challenge has been how to estimate with more
11 certainty the level of need for the purpose of assessing
12 program size and design. Having more definitive data on the
13 type of unmet need could also inform policy discussions
14 related to programs that serve to provide direct grant
15 assistance or programs that reduce energy use, such as energy
16 efficiency or energy conservation education.
17 In 2009, Avista commissioned a Study by the Institute for
18 Public Policy and Economic Analysis at Eastern Washington
19 University. This is attached as Exhibit No.7, Schedule 2. The
20 purpose of the study was "Assessing Heating Assistance
21 Programs in Spokane County". 2 Even though this study was
2 "Assessing Heating Assistace Progr in Spokae County, Intitute for Public Policy & Ecnomic Anysis
(Grt Forsyt, PhD, D. Patrck Jones, PhD, and Mark Wagner). Janua 2010.
KopczYnski, Di
Avista Corp
Page 15
1 limited to Spokane County, we believe the results may have
2 application to other parts of our service area.
3 As noted in that report, the study examined "the recent
4 experience of the two largest heating assistance programs in
5 Spokane County: the federal Low Income Home Energy Assistance
6 Program (LIHEAP) and the Avista Utilities' tariff-funded Low
7 Income Rate Assistance Program (LIRAP) in Washington.The
8 study's central goal was to assess the reach of these programs
9 among the eligible population.
"3 The study provided the
10 following key findings:
11 1. The average heating burden (heating costs divided by
12 total household income) for a household in the US is
13 1. 3%.4
14 2. The average heating burden for households in Spokane
15 County is 1.4%, very close to the US average. 5
16 3. The average gross heating burden for low-income
17 customers (defined as those customers assisted by
18 Spokane Neighborhood Action Programs, or SNAP, which
19 uses the 125% of the federal poverty guideline) is
20 6.1%. 6
3 id., Page 1
4 id., Page 2
Sid., Page 2
6 id., Page 3
KopczYnski, Di
Avista Corp
Page 16
1 4. The average net heating burden for low-income customers
2 is 1.4% (net being defined as heating costs less energy
3 grants, divided by total income). ~ 7
4 5. In 2009, the report shows that 30% of eligible
5 households were assisted by SNAP. This is much higher
6 than the national average of 16%.8
7
8 In short, this report demonstrates that limited income
9 customers served by SNAP have a net energy (heating) burden
10 that is not much different than the average household in
11 Spokane County.
12 Q.Bow will the results of this Study be distributed
13 aDd used?
14 A.The study results are being provided to
15 organizations and individuals that have involvement and
16 interest in energy assistance or energy efficiency programs
17 for these population sectors. Organizations include Community
18 Action Agencies, State and Federal legislators, low-income and
19 senior advocate organizations,and other interested
20 organizations.
21 Q.Wht is the Company doing to help customers mage
22 their energy bills?
7 id., Page 3
8 id., Page 3
KopczYnski, Di
Avista Corp
Page 17
1 A.In addition to the many efforts the Company has made
2 to control costs and improve operating efficiencies, the
3 Company works hard to build lasting ways to help customers in
4 managing their energy bills. Avista is committed to reducing
5 the burden of energy prices for our customers most affected by
6 rising energy prices, including low income individuals and
7 families, seniors, disabled and vulnerable customers.To
8 increase our customers' ability to pay, the Company focuses on
9 actions and programs in four primary areas:1) advocacy for
10 energy assistance programs providing direct financial
11 assistance; 2) low income and senior outreach programs; 3)
12 energy efficiency and energy conservation education and 4)
13 support of community programs that increase customers' ability
14 to pay basic costs of living. The following are examples of
15 these outreach programs to customers:
16 · Gatekeepers Program: Avista has implemented the17 Gatekeepers Program, a program that trains field18 personnel to be aware of signs that a customer may be19 having difficulty with daily living tasks (e.g., paper or20 mail not collected, disheveled appearance, etc). The
21 CARES representatives conducted training of company-wide
22 field personnel who come into contact with residential23 customers on a regular basis. In the event employees24 identify a customer having difficulty, the employee is
25 asked to notify the CARES representatives who would26 contact appropriate community resources for assistance.
27
28 · Senior Energy Outreach: Avista has developed specific29 strategic outreach efforts to reach our more vulnerable
KopczYnski, Di
Avista Corp
Page 18
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
customers
efficiencysafety.
(seniors and disabled customers)
information that emphasizes
with energy
comfort and
.Senior Publications: Avista has created a one-page
advertisement that has been placed in senior resource
directories and targeted senior publications to reach
seniors with information about energy efficiency, Comfort
Level Billing, Avista CARES and energy assistance. A
brochure with the same information has also been created
for distribution through senior meal delivery programs
and other senior home-care programs.
.Power to Conserve: In partnership with KRE television,
a half-hour television program is anually developed that
covers low-cost and no-cost ways to save energy at home.
The goal of the program is to help limi ted income seniors
and other vulnerable populations with their energy bills
by providing home energy conservation education. Theprogram provides helpful energy conservation tips,
information on community resources and ways for customers
to manage their energy bills. A DVO of the program has
also been produced which is included as part of energy
conservation kits provided in senior conservation
workshops.
.Every Little Bit House: In partnership with KREM
television, the long-running "Power to Conserve" program
was updated to profile energy efficiency work done on an
actual Avista customer's home utilizing the low incomeweatherization program provided by SNAP. The program
utilizes a series of commercial vignettes that are
specifically tàrgeted to provide helpful energy
conservation tips, information on community resources and
ways for customers to manage their energy bills. Its
primary target audience is limited income, senior and
vulnerable customers.
Q.Please describe Avista utilities'demd-side
41 management (DSM), or energy efficiency programs.
Kopczynski, Di
Avista Corp
Page 19
1 A.The Company's innovative Energy Efficiency Tariff
2 Rider approved by the Commission was the country's first
3 distribution charge to fund DSM, and is now replicated in many
4 other states. It has provided consistent funding for the
5 delivery of energy efficiency services. Mr. Folsom provides
6 more detail about Avista Utilities'extensive energy
7 efficiency services.
8 Q.Please describe the recent results of the Comany's
9 Project Share efforts?
10 A.Project Share is a community-funded program Avista
11 sponsors to provide one-time emergency support to families in
12 the Company's region. Avista customers and shareholders help
13 support the fund with voluntary contributions that are
14 distributed through local community action agencies to
15 cus tomers in need.Grants are available to those in need
16 wi thout regard to their heating source. Avista Utilities'
17 customers donated $302,300 on a system basis in 2009, of
which
18 $81,700 was directed to Idaho Community Action Agencies. In
19 addition, the Company contributed $111,800 to Idaho customers
20 in 2009.
21 Q.Does the Comau offer a bill-assistance program?
22 A.Yes.In these challenging times, more customers
23 have been finding it more difficult to pay their monthly
KopczYnski, Di
Avista Corp
Page 20
1 energy .bill.In an effort to assist and educate customers
2 about options such as Comfort Level Billing,Payment
3 Arrangements, and Preferred Due Date, we developed a campaign
4 encouraging customers to learn about and enroll in the various
5 bill assistance options available to them. This campaign was
6 launched in March 2009 in both Idaho and Washington. It
7 explains how Comfort Level Billing helps smooth out the
8 seasonal highs and lows of customers' energy usage and
9 provides the customer the option to pay the same bill amount
10 each month of the year. This allows customers to more easily
11 budget for energy bills and avoid higher winter bills.This
12 program has been well-received by participating customers.
13 Over 19,187, or 14%, of Idaho electric and natural gas
14 customers are on Comfort Level Billing.
15 In addition, the Company's Contact Center Representatives
16 work with customers to set up payment arrangements to pay
17 energy bills, and choose a preferred due date.In 2009,
18 35,459 Idaho customers were provided with over 89,092 such
19 payment arrangements.
20 Q.Please sumrize Avista's CARS program.
21 A.In Idaho, Avista is currently working with over
22 1,116 special needs customers in the CARES program. Specially-
23 trained representatives provide referrals to area agencies and
KopczYnski, Di
Avista Corp
Page 21
1 churches for customers with special needs for help with
2 housing, utilities, medical assistance, etc.
3 Q.Bave these programs helped mitigate the impact on
4 customers in need?
5 A.Yes. In the 2008/2009 heating season, 9,788 Idaho
6 customers received $3,740,765 in various forms of energy
7 assistance (Federal LIHEAP program, Project Share, and local
8 community funds). On Septemer 30, 2008, President Bush signed.
9 legislation that provided $5.1 billion for the Low Income Home
10 Energy Assistance Program (LIHEAP) for the 2008/2009 heating
11 season. This increased funding was to serve an additional 2
12 million households and raise the average grant from $355 to
13 $550 and also allow states to carryover any funds remaining to
14 the next year's heating season.Idaho's share of the LIHEAP
15 funding was increased from $12,376,000 to $26,940,000.
16 On December 16, 2009, President Obama signed an omnibus
17 appropriations bill that continued to provide $5.1 billion in
18 funding for the Low Income Home Energy Assistance program for
19 the current fiscal year. The LIHEAP funding includes $4.5
20 billion in formula funds and $590 million in contingency
21 funding. LIHEAP and many other government programs had been
22 operating under funding provided through a continuing
23 resolution that was set to expire December 18, 2009.Idaho's
KopczYnski, Di
Avista Corp
Page 22
1 share of the LIHEAP funding was increased from $26,940.000 to
2 $28,094.000. This bill also provided increased funding for
3 weatherization assistance programs. These programs and the
4 partnerships we have formed have been invaluable to customers
5 who often have nowhere else to go for help.
6 Q.Can you please describe how the Comany measures
7 customer satisfaction, and how important it is to Avista?
8 A.Yes, our customer satisfaction is very important to
9 Avista. We measure satisfaction by doing a quarterly survey we
10 refer to as the "Voice of the Customer" (VOC). The purpose of
11 the VOC Survey is to measure and track customer satisfaction
12 for Avista Utili ties' "contact" customers - customers who have
13 contact with Avista through the Call Center and/or work
14 performed through an Avista construction office.Avista
15 Utilities' company goal for customer satisfaction is measured
16 by thi s Survey.
17 Customers are asked to rate the importance of several key
18 service attributes.They are then asked to rate Avista' s
19 performance with respect to the same attributes (time for
20 connection to a representative, representative being courteous
21 and friendly,representative being knowledgeable,being
22 informed of job status, leaving property in condition found,
23 etc. )Customers are also asked to rate their satisfaction
KopczYnski, Di
Avista Corp
Page 23
1 with the overall service received from Avista Utilities.
2 Customer verbatim comments are also captured and recorded.
3 Our most recent fourth quarter 2009 customer survey
4 results show an overall customer satisfaction rating of 94% in
5 our Idaho, Washington, and Oregon operating divisions.This
6 rating reflects a positive experience for the vast majority of
7 customers who have contacted Avista related to the customer
8 service they received.
9
10 Q.Does this conclude your pre-filed direct testimony?
11 A.Yes.
KopczYnski, Di
Avista Corp
Page 24
DAVID J. MEYER
VICE PRESIDENT AN CHIEF COUNSEL OF
REGULATORY & GOVERNNTAL AFFAIRS
AVISTA CORPORATION
P . 0 . BOX 3727
1411 EAST MISSION AVENUE
SPOKAE, WASHINGTON 99220-3727
TELEPHONE: [phone]
FACSIMILE: [phone]
DAVID. MEYER~AVISTACORP. COM
BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION
IN THE MATTER OF THE APPLICATION ) CASE NO. AVU-E-10-01
OF AVISTA CORPORATION FOR THE ) CASE NO. AVU-G-10-01
AUTHORITY TO INCREASE ITS RATES )
AN CHAGES FOR ELECTRIC AN )
NATURA GAS SERVICE TO ELECTRIC ) EXHIBIT NO. 7
AN NATURA GAS CUSTOMERS IN THE )STATE OF IDAHO ) DON F. KOPZCYNSKI
)
FOR AVISTA CORPORATION
(ELECTRIC AN GAS)
Customer Usage
State of Idaho.. Electric & Gas
As of December 31, 2009
Electric kwh
Schedule No. of Customers (OOOs)% of Total kwh
Residential Sch. 1 100,073 1,182,368 34%
Genera Sch. 11&12 19,420 322,570 9%
Lge. General Sch. 21 &22 1,418 699,953 20%
Ex. Lge. General Sch. 25&25P 9 1,158,336 34%
Pumping Sch. 31&32 1,315 58,885 2%
Street & Area Lights 123 13,816 0%
122,358 3,435,927 100%
Natural Gas Therms
Schedule No. of Customers (OOOs)% of Total Therms
General Servce 101 72,939 56,909 45%
Lg. General Servce 111&112 1,057 19,554 16%
High Anual Load 121&122 0%
Interrptible Servce 132 437 0%
Transporttion Service & Other 9 48,773 39%
74,006 125,672 lÔO%
196,364Total Electrc & Gas Customers
Exhibit No.7
Case No. AVU-E-I0-0l and AVU-G-I0-0l
D. Kopczyki, Avista
Schedule 1, p. 1 of 1
Institute for Public Policy &
Economic Analysis
Assessing Heating Asistnce
Programs in Spokane Count
.
By:
Grant Forsh, Ph.D.
D. Patrick Jones, Ph.D.
Mark Wagner, MA
.
...
.
January, 2010
.
.
Exib No.7ca No. AVl-1()1 an AYU1()1D. Ko, Avi
SCul 2. Pa 1 of 61
Assessing Heating Assistance Programs in Spokane County
By
Grant Forsyth, Ph.D.
Department of Economics
Institute for Public Policy and Economic Analysis, EWU
D. Patrick Jones, Ph.D.
Institute for Public Policy and Economic Analysis, EWU
Mark Wagner, MA
Institute for Public Policy & Economic Analysis, EWU
A Report to
Avista Utilties, Spokane, Washington
January, 2010
Exbi No.7
Ca No. AVU-1G-1 an AVt1G-1D. Ko, Avi
Scul 2. Pa 2 of 61
Table of Contents
List of Tables.............................................................................................................................................................11
List of Figures...........................................................................................................................................................iii
Acknowledgements...................................................................................................................................................iv
1. Executive Summary............................................................................................................................................1
2. Study Origins .......................................................................................................................................................4
3. Program Description and Definitions...............................................................................................................5
4. Data, Methods and Organization of Analysis................................................................................................10
5. Estimation of At Risk Households in Spkane County.................................................................................12
6. Analysis of Spokane County Households Asisted by L1HEA & L1RA......................................................25
7. Measuring Heating Expenditure Shares for All of Spokane County..........................................................35
8. caveats, Qualifications, Conclusions.............................................................................................................48
References................................................................................................................................................................50
Appendix A: Key to City and Town Abbreviations..............................................................................................51
Appendix B: Heating Shares for Spokane County census Tract in Heating season 2008...........................52
Endnotes...................................................................................................................................................................55
Exbi No.7ca No. Avu-1G-1 an AYU1G-1D. Ko, AviSc 2. Pa 3 of 61
List of Tables
Table 3.1: Recent National, Washington State & Spokane County L1HEA Allocations.............................5
Table 3.2: Home Heating Shares, U.S. and Western U.S. for Federal Fiscal Year 200.................................6
Table 5.1: Projecions of At Risk Households in Spokane County for 2009-2012...........................................24
Table 6.1: Analysis of SNAP (L1HEAP+L1RA) Households, Heating Season 200..........................................26
Table 6.2: Analysis of SNAP (L1HEAP+L1RA) Households, Heating Season 2008..........................................27
Table 6.3: Analysis of SNAP (L1HEAP+L1RAP) Households, Heating Season 200. ........................................28
Table 6.4: Growh Analysis of SNAP (L1HEAP+L1RAP) Households, Heating seasons 200
To 2009...................................................................................................................................................................30
Table 7.1: Average Residential Heating Costs by Fuel Type............................................................................36
Table 7.2: Summary of the Frequency Heating Shares in Spokane County, Heating Season 200............38
ii
Ex No.7
Ca No. AW-E-1D-1 an Avu1D-1D. Ko, AvISC 2. Pa 4 of 61
Ust of Figures
Figure 3.1: Average U.S. Household Expenditures on Energ by Income Quintile, as a Share of
Household Income................................................................................................................................................. 7
Figure 5.1: At-risk Households by Povert Adjustent in 1999, 2003, and 2008........................................15
Figure 5.2: At Risk Households by Census Tract at the 125/150% Adjustment in 2008..............................16
Figure 5.3: Estimates of the At Risk Householders 65 and Over in 1999, 2003, and 2008.........................17
Figure 5.4: 1999 Relationship between At-risk Households & At-risk Householders 65 and Over
at the 125/150%djustment..................................................................................................................................17
Figure 5.5: SNAP lIHEAP+lIRAP Households frm 2002008/2009.............................................................18
Figure 5.6: Regression Relationship between ARH and Median Household Income across
39 Washington Counties in 1999.........................................................................................................................19
Figure 5.7: SNAP lIHEAP+lIRAP Households from 200-200........................................................................21
Figure 5.8: Regression Relationship betwen ARH and Median Household Income across 39
Washington Counties in 1999..............................................................................................................................23
Figure 6.1: Cumulative Freuency Distribution of Heating Burden of SNAP Households in Heating
Seasons 200, 2008 & 200.................................................................................................................................32
Figure 6.2: Distribution of SNAP (lIHEAP+lIRAP) Households by Zip Code in Heating seasons
2002009..............................................................................................................................................................33
Figure 6.3: Analysis of SNAP (lIHEAP+lIRA) Households, Heating Season 200.......................................34
Figure 7.1: Average Heating Cost by census Tract for Spokane Count.....................................................40
Figure 7.2 Average Heating Cost by census Tract for the City of Spokane .............................................41
Figure 7.3: Median Income by Census Tract for Spokane County................................................................43
Figure 7.4: Median Income by census Tract for City of Spokane..................................................................44
Figure 7.5: Heat Burden by Census Tract for Spokane County......................................................................46
Figure 7.6: Heat Burden by Census Tract for City of Spokane........................................................................47
ii
Ex No7Ca No. Avu-1G-1 an AYU1G-1D. Ko, AviSC 2, Pa 5 of 61
Acknowledgements
This project would not have been possible without the swift and able assistance from seeral Spokane
County utilties and SNAP. In particular, would like to thank Allen Cousins, senior GIS analyt, and Rob
Wagner, business analyt, both at Avist Utilities, for their GIS expertise and data sets, respectively.
Gene Steinolfsn, member service manager at Inland Power and Ught, provided the research team with
biling data. Joe Noland, light department director at the City of Cheney, also provided us with billing
data. Wai landry, comptroller at Modem Electric Water Company, in Spokane Valley, gave us data totals
that assisted the analysis. Finally, we want to thank the SNAP staff, particularly Alice Damm, for the
many hours spent with the research team.
We also would like to acknowledge Anne Marie Axorthy and Christine McCbe, for their strategic input
and guidance during the cours of the project. None is, of course, responsible for any errors of omission
and calculation on our part.
iv
Ex No.7ca No. AVl-10.1 an AYU10.1D. Koki, Avi
Scul 2, Pag 6 of 61
... :-.......;~~
1
Ex No.7Case No. AVl-1()1 an AVl1()1D. Ko, Avi
Sc 2. Pag 7 of 61
2
Eilbi NO.7ca No. Avu-1D-1 an AVl1D-1D. Ko, AvI
SCul 2, Pag 8 of 61
3
Exib No.7
ca No. AW-E-1G-1 and AVl1G-1D. Ko, Avi
SCule 2, Pa 9 of 61
2. Study Origins
Low-income residents in Spokane County
are eligible for two kinds of financial
support for their heating needs. The first
comes from a federal program, the Low Income
Home Energy Asistnce Program, or UHEA.
Originally enacted in 1981, its current purpose
is "to assist low income households, particularly
those with the lowest income, that pay a high
proportion of household income for home
energy, primarily in meeting their immediate
energy needs." 1 A second source of relief
comes from the large investor-owned utlity in
the eastern Washington, Avista. Since 2001, it
has funded a similar program to L1HEAP, the
Low Income Rate Asistance Program, or L1RAP.
In heating year 200-2009, 10,459 households
in the County received L1HEAP assistance. In the
same heating year, 2,681 County households
were able to take advantage of L1RA, for a
combine total of 13,140 households assisted.
This represented an increase of nearly 4,00
households aided by the two programs from
the prior year, largely due to monies put into
the L1HEAP program by the federal American
Recovery and Reinvestment Act.
Despite this impressiv jump in coverage,
administrators, users and funders of the two
programs are concerned about the programs'
adequacy in covering all Spokane households
eligible for heating assistance. As a
consequence, the Institute for Public Policy &
Economic Analysis at Eastern Washington
University was commissioned to study the
issue. Specifically, the Institute was charged
with investigating:
.The definition of energ burden for low-
income households;
An estimate of the total number of low-
income households in the County who
currently qualif for one of the two
programs under some definition of energ
burden;
.
· An estimate of the number of low-income
households in the County who willilcly
qualify over a subsequent three year
period;
· An estimate of the number of low-income
households headed by seniors who
currently qualif for the two programs
· An analysis of the households recently
served by the two programs; and
· A depiction of the geographical distribution
of households served by the two programs
and households who generally might
qualify.
4
Ex No.7
ca No. Aw--1G-1 an AYU1G-1D. Ko, Avi
SC 2, Pag 10 of 61
3.Program Review & Definitions
3.1 UHEAP
LIHEAP currently targets two types of low
income households: those with high
"burden" and those who are "vulnerable".
High burden is generally defined as very low
incomes and high home energy costs, while
vulnerable households consists of those with at
least one young child ((: 5 years), or a member
over 60 years of age, or a member with
disabilties. The federal L1HEAP statute defines a
low income household as one at or below the
150% federal povert level (FPL) or the 60
threshold of a state's median household
income, whichever is greater.
As a block grant program, UHEAP's eligibilty
standards vary by state. Since the federal
dollars allocated to each state are inadequate
to cover all households who qualify, most states
use the FPL threshold, a lower amount than the
median income measure. Federal statute allows
states to set a threshold below the 150% of the
FPL, but it must lie above 110% of the FPL. In
Washington State, the administrator of the
program, the Department of Commerce, uses
the threshold of 125% of the FPL. 2 In Spokane
County, L1HEAP is administered by the Spokane
Neighborhood Action Program, or SNAP, a
community action agency in operation since
1966. The relative sizes of the two most recent
L1HEAP allocations are shown in Table 3.1.
In heating season 2007-oS, Washington State
received 2% of all federalllHEAP dollars; in
heating season 200-0, the share slipped to
1.7%. While the state's population in 200S
made up 2.2% of the U.S. total, it estimated
povert rate (at the FPL), at 11.3%, was
considerably lower than for the U.S., estimated
at 13.2%. In heating season 2007-0, Spokane
County received S.2% of the state total; in
heating season 200S-0, the County share was
S.O%. While Spokane County made up 7% of the
state's population in 200S, its estimated
Table 3.1: Reent National, Washlna Stae
& Spkane County UHEAP Allotions
Juriict
U.S 1,977,027,460 4,476,301,613
WA a 40,449,571 74,602,937
Spokane
County b 3,323,914 5,993,070
a. The Washington state allocation included
$1.631M and $3.035M for tribal governments in
the two years.
b. Spokane County values are actual expenditures
Sources: for the U.S. & Washington: U.s.
Departent of Health & Human Servces,
Administrtion for Children & Families,
http://www.acfhhs.qov/oroqrams/oclliheoo/fundi
nqlfund.html,; for Spokane County, adminisrative
data from SNAP.
povert rate was considerably higher than the
state's: 13.7% vs. 11.3%.3
Although the lIHEAP statute defines assistance
for energy, SNAP administers its program for
the heating season only. This conforms to the
Washington State Department of Commerce
guidelines. As a consequence, this study
examines heating assistance and burden.
3.2 URAP
The Low Income Rate Assistance Program, or
lIRAP, is funded entirely by Avista Utilties and
supplements the assistance offere by L1HEA.
It is meant to extend the reach of lIHEAP;
consequently, if a household receive L1HEAP
dollars, it is ineligible for L1RAP help. The
general eligibilit requirements are the same as
L1HEAP, with apparent preference given to
those with the highest heating burden. The
program is restricted to Avista's customers,
those who are "least able to pay their bils.'14
it is offered in two of the three states that
comprise Avista's service terrtory. Its funding
comes largely from a surcharge on its
5
Exibi No.7
Ca No. Avu-1Ð-1 an AW-G1Ð-1D. Ko. AvlSC 2, Pa 11 of61
customers bils, amounting to approximately
O.S% of the base rates for both electricit and
natural gas. Some funding also comes from the
utiltys philanthropic campaign, Project Share,
as well as from a separate set of donations from
Avista employees and shareholders. In Spokane
County, lIRA expenditures for the 2oo7-oS
heating season amounted to $1,322,496; in the
2OOS-0 seasons, they increased to $1,616,643.
Avista Utilties engages the same community
action agencies who manage the lIHEAP to
administer lIRAP.ln Spokane County, this is
SNAP.
3.3 Determination of Heating Burden
As conventionally defined, heating burden is
generally the ratio of household heating costs
to household income. As such, for an individual
household, the ratio defines the share of total
income taken up by heating expenditures. In
the economics of consumption, analysts
examine this ratio simply as a share, not a
"burden". The latter term implies a position in a
household's budget that creates problems of
matching income with expenditure. Problems ofUmeeting budget" might arise for households,
but likely not low levels of this share. What
then, constitutes a high level, or one that might
be construed as a burden?
One might look at national or regional summary
data to gain some insights. The most recent
detailed, household-level information set at the
national level comes from the Department of
Energys quadrennial Residential Energ
Consumption Survey, or RECS. A summary of
the findings, as reported in the 200 LlHEAP
Home Energy Notebook (200) are presented
below. The data stem from the year of the most
recent survey, 2001, and have ben updated by
the report to 2006 values.
Table 3.2: Hom Heang Share, u.s. and
Weser u.s. fo Feeral Fisl Year 200
Mean IrouP shares
All households 1.1
Low income. 3.S
lIHEAP recipients 6.S
0.6
2.0
3.6
Mean individual
All households
Low income
lIHEAP recpients
2.9
6.3
11.2
1.6
3.3
6.5
Median individualAll households 1.3 O.S
Low income 3.0 1.6
lIHEAP recipients 7.1 5.5
a. Lo Income households are those that fall into
the L1HEA defnition of at or below the 150
threhold of the FPL or at or below 60 of
the stte's median household income.
Sourc: U.S. Departent of Helt & Human
services, Administration for Children & Families
(August, 200), Tabla A-Sa-.
Table 3.2 presents heatini shares in seeral
ways, and these merit a brief discussion. The
first is a distinction between uindividual" and
"group" shares. The former category represents
first the calculation of individual household
ratios or shares, then of the average of these
shares. Group shares are the result of first
summing all individual household heating costs,
then summing all indivdual household
incomes, and dividing total heating cost by
total household income.
The two methods wil typically yield diferent
result, because typically heating cost do not
increase at the same rate (linearly) as incomes
increase. For example, an examination of the
2007 Bureau of Labor Statistics (BLS) Consumer
Expenditure Survey reveals that the relationship
between income and energ (and presumably
heating) expenditures is non-linear, moving
6
Exib No.7
C8 No. Avu-1Q-1 and AVU1Q-1D. Ko. Avl
SCedle 2. Pa 12 of 61
from low- to high-income households, and
highlights the diffculties low income
households face with rising prices or fallng
incomes.
To ilustrate, Figure 3.1 shows the ratio of
average household energy expenditures to
average gross household income for each
quintile in the U.S. in 2007 (Q1 = the poorest
20% of households and 05 = the wealthiest
20%). Notice that for each energ source
(electricity, natural gas, and fuel oils), the
poorest quintile, Q1, always has the highest
relative expenditure share. With respect to
elecricity, Q1's average share is around 8%,
which is twice as high as Q2 and eight times
higher than 05. This means sharp increases in
energ prices will be felt more acutely by low-
incme households since they wil face more
painful reductions (or eliminations) in the
consumption of both energy and other
goods/services to offet the increased share of
energy cost in the household budget. If
households are already operating at the
minimum level of energy use for a livable
environment, then expnditures reductions wil
come entirely from all non-energ related
goods and services.
Figure 3.1: Average U.S. Household Expnditures on Energ by Incom Quintlle, as a Share of
Household Income beore Taxes
9%
I 8%-""..7%i "8'
J 5'"-"-'I 4%
!3%~
l!e.--s 2'
i --1%
0'
Q1 Q2 Q3 Q4 Q5
Qulnle
I --EIe . __Na G_--FueOIIa I
Source: 2007 BLS Consumer Expenditure Sutvy
The pattern in Figure 3.1 is also consistent with
survey data in the 2006 LlHEAP Home Energy
Notebook. The study finds a 200 median
residential energy (all uses, not just space
heating) share of 3.1% higher-income
households, 9.5% for low-income households,
and 15.3% for households that received L1HEAP
support (p. 4, Table 2-1).
Consequently, when one discusses levels of
energy shares, or burden, it is important to note
how the calculation was made - on the basis of
individual household ratios or summing heating
bils and household income for a population
over a known geography, then calculating a
ratio. In this study, both approaches are taken.
Note that Table 3.2 shows both mean, or
average, and median values. (Te median is the
value in a distribution of numbers at which 50%
of the values lie aboe it and 50% of the value
lie below it.) Both are measures of "central
tendency," of the middle of a distnbution of
measurements. Both are measures of what
might be consider "typical". In a symetncl
7
Ex No.7
ca No. AVU-E-10-1 an AVU-G10-1D. Ko, Avi
SCle 2, Pa 13 of 61
distribution, the mean and the median are the
same. In a skewed distribution, they are not. As
Figure 3.1 shows, the distribution of energy
shares is highly skewed. When a distribution is
not symmetrical, very often the median is a
preferred measure of the middle. Finally, Table
3.2 presents the results for both the U.S. and
the Western U.S. Census region.s As one can
observe, considerable differences exist between
the two columns. The Western U.S. shows
lower values than the all-state average.
Several observations follow from Table 3.2. As
stated generally above, the values of mean
heating shares calculated for individual
households are larger than the mean heating
shares calculated on a group basis. Second,
heating shares in the Western U.S. are
considerably lower than the U.S. average,
nearly 50% in most categories. Third, as seen in
the energy shares of Figure 3.1, heating shares
faced by low income households are larger than
for the entire population of households,
typically by more than 100. Fourth, from a
program evaluation perspecive, the much
higher heating shares shown by L1HEAP
recipients reflecs the preference given to the
"Iowest of the low" by most local program
administrators. Fifth, as mentioned above,
median heating shares, calculated on an
individual household basis, are considerably
lower than mean heating shares calculated on
an individual household basis. This stems from
the asymmetrical distribution of heating shares.
Table 3.2 serves as a reminder of the complexit
of measuring heating shares and of the care one
must exercise in setting up benchmarks. Its
values represent the current best measurement
of the size of household budgets taken up by
heating cost. The unanswered question from
these share calculations is at what threshold do
they represent a "burden"?
There are no hard and fast rules to determine
this and one necessarily enters into the realm of
value judgments. In an earlier study (APPRISE,
2005), the authors of the LlHEAP Home Energy
Notebook for 200 discusse three general
approaches to determining a burden threshold.
One involves ordering all households by energy
(or heating) shares and setting a cut-off at a
certain percntage of all households, one that
ostensibly captures the highest burdened
households. This raises the question of where
cut-off should be drawn. Another approach is
to use the statistical tool of standard deviation
and set the cut-off at one stndard deviation
above the mean share value.6 This rule certainly
does not have any rationale beyond the
presumption that households with energy
(heating) shares that are a certain distance
away frm the mean, or typical, household,
deserve some kind of assistnce.
The approach the authors recommend and use
is a variant of the income share approach
depicted in Table 3.2. In a third study by
APPRISE (2007), one for Washington State, they
note the approach taken by the consulting
group Fisher, Sheehan & Colton for enery
burden. This group draws on the literature of
shelter (housing and energ) affordabilit,
which often uses 30% of household income to
set the threshold. Fisher et al then invoke their
own research on energy cost as a share of total
shelter cost to suggest that about 20% is
average. Consequently, the level of energ
share of income at which a burden wil arise is
6% (20% of 30%).
To translate the energy calculations into a
heating threshold, one would need to apply the
percentage of total energ costs taken by space
heating. This varies acros the country.
According to the 2006 LlHEAP Energy Notebook
(2008), space heating takes up about 37% of
total energy costs nationally. Applying this
percentage implies a .burden" threshold of
2.2% for all households.
The Apprise authors use a similar approach.
They note that a .seere" shelter burden is one
in which 50% or more of household income
goes to shelter expenditures. They cite their
own research that 22% of shelter cost are
8
Ex NO.7
case No. AYUE-1()1 an Avu1()1D. Ko, Avl
SCed 2, Pag 14 of 61
attributable to energy expenditures in low
income households. Conseuently, a "severe"
energy burden threshold for these households
is about 11%. They provide similar calculations
for the 30% of household income rule, and
arrive at a value similar to Fisher et al of 6.5%.
They label this a "moderate" residential energy
burden.
To translate these two cut-off points in
household income into a heating burden
threshold, they apply a 39.3% share taken by
heating and cooling nationally of energ
expenditures. The results: "high" heating It
cooling burden is 4.3%, while a "moderae"
heating It cooling burden is one greater than
2.6% but less than 4.3% of household income.
Compare these thresholds to the average values
reported in Table 3.2 for Western U.S. low
income households: 2.0% for the group
calculation and 3.3% for the individual
household calculation.
As noted in the introduction, this study
examines heat burden. The focus on heat
burden reflect the distribution of L1HEAP and
lIRAP monies over the winter months in
Spokane County. The distribution of monies for
winter heating bils is a response to regional
energ bils spiking during the coldest (rather
than the hottest) months of the year.
Therefore, if the heat burden of Spokane
County households exceeds the lIHEAP
thresholds for both heating and cooling costs,
then it is likely that their energ burden for
heating and cooling is higher.
While threshold calculations are necessary to
arrive at some operational rules for evaluating
low income heating programs, it is obvious that
they rest on certain assumptions. Whether
these are correc is the subject of ongoing
research. It bears repeating that these
calculations are based on national averages. As
table 3.2 makes clear, there are distinc regional
variations. Indeed, as the Apprise study for
Washington state (2007) notes, there are
substantial differences in energy costs, and
presumably energy (as well as heating)
thresholds within the state.
Finally, these national averages obscure
variations among groups targeted by low-
income heating assistnce programs. For
example, it is likely that the senior population
has an expenditure mix different than the
population at large. If they are home owners
and have been living In the same dwellng for
years, they, as a group, may face lower shelter
costs as a share of. their income, since the home
may be paid for. On the other hand, medical
expenses may take a much higher share of
household Income, especially for the older
seniors.
9
Ei No.7
ca No. Avu-1()1 an AVU-G1()1D. Ko, Avi
Scle 2, Pag 15 of 61
4. Data, Methods & Organization of
Analysis
To directly examine heating burdens for
Spokane County, one needs accurate
heating costs and income data for each
household. Ideally, this would be provided by a
household census; however, a representative
sample, such as the RECS, would work. The
research team did not have access to either
tool. An alternate, less detailed method uses
heating cost and/or income data at the lowest
geographical unit possible, following the
"group" approach discussed in secion 3.2. This
method was employed in two variants for this
study.
The first looks at Census income data. The
smallest unit for which income data could be
secured was the census tract.7 The goal of this
approach is to produce a current snapshot of
the distribution of household income for each
census tract in the County. In this way,
estimates of the number of households below
certain income levels can be developed. These
numbers, for those "at risk," are viewed as
equivlent to the number of households facing
an energy burden. As noted by the APPRISE
(2005) study, "Households with incomes less
than $20,00 per year represent over 95% of all
households that have a high home energy
burden. Almost two thirds of households with
incomes below $10,00 are characterized as
having a high home energy burden." i In other
words, if one can determine the number of low
income households, one has a fairly accurate
estimate of households facing heating burdens.
To complete this analysis for Spokane County,
the research team used income data from the
200 Census (actually 1999) as the base.
Income levels for all households in each tract
are then "brought forward" to the present via
the techniques described in section 5.1. The
result is a current estimate of households who
qualify for lIHEAP or lIRA assistance. A by-
product of the estimate of total at-risk
households is an estimate of the number of at-
risk households with at least one member age
65 or over. This is taken up in section 5.3.
The extension of this technique to the near
future is taken up in section 5.5. Techniques
employed in this section are largely those of
extrapolation of historical trends, in both a
linear and non-linear way.
The second variant of the group approach
tackles the creation of average heating costs in
census tract. This necessitated securing source
data on heating cost from the Countýs elecric
and gas utilities. The research team was able to
do this with data from three of the five utilities,
representing the vast majority of households.
The calculation, however, of heating cost frm
these records was hardly straightforward. Firs,
households may use two utilities, one for
electricity and one for natural gas (Avista), but
without the abilit to match addresses, we
could not identif them and calculate only
heating costs. Send, households may use one
utilty for electricity but heat with fuel oil or
propane gas. Since the research team had no
source data from fuel oil or propane gas
providers, we face a similar inabilty to match
records. As a result, census tract average
electricity cost information from utilities other
than Avista had to be adjusted to account for
these "dual" utilty households.
In the end, we were able to fashion a version of
a "group" measure of energ share or burden
for each census tract for 200. (In the study,
heating seasons are labeled by the year in
which they end.) We emphasize, however, that
this method does not yield the number of
households in each census tract that face a
heating burden, since we could not line up
heating cost records with an income
distribution. Further detail about the method is
taken up in section 7.
Thanks to thorough and clean records kept by
SNAP, the research team was able to analye
data for the subst of Spokane County "at risk"
10
Ex NO.7ca No. AW--10-1 an AVU10-1D. Ko, Avi
SCle 2, Pa 16 of 61
households who have recently been served by
SNAP. The results are characterized in seion 6.
This analysis, in contrast to the proxy
techniques employed in other sections, yielded
unambiguous burden data for three heating
seasons: 200, 200 and 200. Of particular
note is the calculation of pre- and post-award
heating burdens for households receiving SNAP
assistance. section 6 also displays the
distribution of SNAP awards, by level of energy
burden and zip code. Combined with the
estimates of at-risk households in section 5, the
SNAP numbers give a sense of the size of
"unmet nee," or of the number of eligible
households who have not received heating
assistnce.
The final chapter considers the assumptions
necessarily employe in the analysis, as well as
the limitations of both methods and data. It
concludes with a brief discussion on the validity
of the studýs estimates.
11
exib No.7case No. Avu-1G-1 an Avu1G-1D. Ko, Avi
SCule 2, Pa 17 of 61
5. Estimating At-risk Households
Households with a high probabilty of
qualifying for energ assistance funds
are defined in this study as "at-risk
households" (ARH). Under the current L1HEAP
program, the income threshold for eligibilty is
set at 150% of the appropriate federal povert
level (the povert level applied to the 150%
adjustment is dependent on housèhold size) or
60 of a state's median household income,
whichever is higher.
5.1 Methodloe for Esmating At-risk
Housholds
The most complete data on the distribution of
household income (HHI) by census tract comes
from the 200 census, which uses 16 income
brackets for sorting occupied households by
their 1999 HHI. Therefore, to generate annual
estimates for the 2003-2008 periods, the 1999
share of total households in each income
bracket for each tract is multiplied by annual
estimates of total households in each tract.
This means annual estimates of total
households per tract for the 2003-2008 period
are allocated over inflation adjusted income
brackets using the bracket shares from the 200
census. This approach assumes that the share
of households in each inflation adjusted income
bracket has not changed significantly since
1999, even though the number of households is
not constant over time.
More formally, the estimation process is:
(1) hl,.t = (Hc,t H$¡,c99)
for i = 1,...,16 income brackets; c = 1,...,106 census
tract; and t = 2003,...,200
Where:
hi.t is the estimated number of occupied
households in Inflation adjusted Income bracket i in
tract c at time t;
He. is the estimate of total occupied households in
tract c at time t; and
5i Is the share of total ocupied households In
Income bracket i in tract c in 199, as reported in the
200 U.S. census.
Therefore, it follows from equation (1):
16 10
(2) Hc:t = Lhi.ct and Zt = LHc:t1-1 em1
Where: It Is the total estimated ocupied
households in Spokane County at time t.
The annual estimates of ocupied households
by tract come from Washington's Ofce of
Financial Management (OFM) for the 2003-
2008 period. In order to adjust for the impact
of inflation, the 1999 income brackets are
increased using the Consumer Price Index (CPI)
for cities in the western U.S. with fewer than
1.5 millon people (the Western blc index).
The adjustment to the 199 brackets for the
years 2003-200 was as follows:
(3) Bi.t= (ßi)(1 + Ft) and BU,l.t= (Bu.99)(l + Ft)
Where:
Bu,t and Bu.u are the lower (l) and upper (U) incme
limit for bracket i in year t;
BW9 and Bu.1 are the lower and upper limit for
bracket I in 199; and
Ft is the totl amount of inflation that has occurrd
beteen 1999 and year t.
A similar approach is used for estimating the
share of at-risk households with a head of
household 65 years or older (ARH65). However,
since the OFM only estimates total households,
an additional variable is added to equation (1)
to estimate those households with a
householder 65 years or older. This variable
(P c,) is the share of 65 and over households in
tract c in 199 from the 200 census. Thus,
equation (1) becomes:
(4) ie= ((Hc,)(Pc,99))P1,c)
for i = 1,...,16 income bracket; c = 1,...,106 tract;
and t = 2003,...,200
12
Exib No.7Ca No. Aw--1G-1 an Aw-1G-1D. Ko, AviSC 2, Pa 18 of 61
Where:
kit is the estimate number of 65 and ovr
occupied households in income bracket i in tract c at
timet;
Hcot is the OFM estimate of total occupied
households in tract c at time t;
P co99 is the share of total 65 and over occupied
households in tract c In 1999; and
PI, is the share of total 65 and over occupied
households in income bracket i in tract c in 199 as
reported in the 200 census.
In other words, the term ((Hcot )(P c,9)) is an
estimate of the total 65 and over households in
tract c at time t. This is then multiptied by the
1999 share of 65 and over households in each
income bracket in tract c to estimate ki.t.9
The next step is to estimate the number of at-
risk households (ARH), using the federal povert
tines (FPL) for each year since 1999. FPL levels
increase as the number of people in a
household does. Since the census does not
report data on individual households, the
average household size is used to estbtish a
povert tine that would apply on average.
Spokane Countys average household size for aU
households was approximately 2.4 people over
the period of interest; consequently, the
analysis uses the average of the thre-persn
and two-person povert levels. LIkewise, since
the average household size with a 65 or older
householder is approximately 1.4 people, a
similar approach is apptied, using the poert
levels for householders 65 and over for one-
person and tw-person households.
These povert tines are then inflated by 125%
and 150%. The 125% adjustment reflect the
current threshold used by SNAP, and the 150%
reflect one of L1HEAP's legislated maximum
thresholds. One additional threshold is
estbtished by applying L1HEAP's alternative
maximum, defined as 60 of Washington's
median household income (HHI). This threshold
applies to aU household tyes. (At the time of
this writing, 2009 data on povert thresholds,
the CPI, and OFM household estimates were
not available; therefore, ARH for 200 could not
be estimated using the approach described
here.)
These adjusted povert lines are then compared
against the income brackets described by (3).
The number of households assciated with
income brackets at or below the adjusted FPL
are then summed to estimate ARH in each
census tract In this approach, the highest
appticable income bracket is the one in which
the adjusted povert line falls.10 Therefore,
i
(5) rc.t = L hi,cot
1-1
for c = 1,...,106 tract; and t = 2003,...,2008
Where:
r cot is the number of at-nsk households (ARH) in tract
c at time t;
I is the number of income bracket at or below the
adjusted povert line in tract c at time t; and
hw is the number of estimated housholds in the
applicable income bracket in tract c at time t.
Use of the average poert level for a 2- and 3-
persn household resulted In both the 125%
and 150% adjusted FPL encompassing the first
three income brackets (I = 3). In contrast, the
60 median HHI adjustment encompassed the
firsflve brackets (I = 5). Summing across aU
tract In each year, the county total of ARH is:
(6)
10
Rt = Lrc.t
c-l
for t = 2003,...,200
Where: Ri is the estimated county total of ARHs.
Likewise, for ARH65:
i
(7) ec.t = L k...t
1-1
for c = 1,...,106 tract and t = 2003,...,200
Where:
ee. is the number of ARH65 In tract c at time t;
I is the number of income bracket at or beow th
adjusted port line In trct c at time t; and
ic is the number of estmated housholds with at
least one member age 65 or over in the applicable
income bracket in tract c at time t.
13
Exib No.7ca No. AYU1Ð-1 and AVl1Ð-1D. Ko. Avi .
SCul 2. Pag 19 of 61
Use of the average povert level for a 1 and 2
person household (with a householder 65 and
over) resulted in both the 125% and 150% FPl
adjustment encompassing the first two income
brackets (I = 2). In contrast, the 60 median
HHI adjustment encompassed the firs jive
brackets (I = 5). Therefore, as before, the
county total of ARH65 would be:
(8)
106
Et = ¿ec,t
c-1
for t = 2003,...,200
Where: Et is the estimated county total of ARH65.
5.2 ARH Estimation Result
Figure 5.1 (Graphs 5.1 and 5.2) shows the
estimation of the number of at-risk households
(ARH) in Spokane County. Graph 5.1 show the
abslute number of ARHs in 1999, 2003, and
2008; Graph 5.2 shows the share of ARHs to
total county households in 1999 and 200. As
of 2008, there were approximately 43,00 ARHs
at the 125/150% FPl adjustment and 69,700 at
the 60 of HHI adjustment. The relatively sharp
jump from 2003 to 2008 reflect a stronger than
normal growth in county households starting in
2005. Although the absolute number of ARHs
has increased since 1999 (7% to 7.5%
depending on the FPL adjustment used), the
estimated shares of ARH have not changed
signifintly since 1999. At-risk households
represent about 24% and 39% of all households
at the 125/150% and 60 adjustments,
respectively.
14
exibi No.7ca No Avu-1()1 and AVl1()1D. Kop, Avi
Sc 2, Pag 20 of 61
Fiiure 5.1: Esmates of th Number of At-rk Houholds by Povert Adjus In 1999, 2003 and
200
80,00
Gr 5.1: Tot AR by Pov Adtm
70,00
60,00
50,00
40,000
-30,00
20,00
10,000
12550 X 2. P_ FP
-199-20-20
60% X WA Me HHI
45%
Graph 5.: Sha of AR to Tot Ho..ho by Pov Ad
40
35%
30%
25%
20%
15%
""
5%
0%
125150 X 2.5 Pe FPL
-199
-203-20
60 X WA Med HH
It is useful to compare these numbers to those
from the 2008 L1HEAP evaluation stdy based
on the 2001 Residential Energ Consumption
Survey (RECS).l1 Using the actual eligibilty
standards used by states, the study found that
about 21% of U.S. households qualified for
L1HEAP assistance. This number is within the
range of the Countýs estimated share of ARH
shown in Figure 5.1.
A further check of this result can be made by
comparing it to Washington State's total low-
income household estimate provided by
Apprise (2007, p. 4). For 2005, the study arrived
at 452,252 households at or below the 150%
FPL. No county break-out was given, however.
Apportioning Spokane Countýs population
share of the 2005 state total (7%) would yield
31,538. However, the County has been
1S
Exbi No.7
ca No. AYU-1G-1 and Avu1G-1D. Kopsk, Avi
Scule 2. Pag 21 of 61
characterized by a higher rate of povert than
the state. In 2005, the all-age povert rate
(100% FPl) in Spokane County was estimated at
14.4% vs. 11.9% for the State, or 21% higher. U
After factoring in this adjustment, the Countys
povert-adjusted population share is 8.4%.
Applying this to the total reported by APPRISE
yields 38,144 households.
While separated by three years, the results
from this study and the one provided by this
derivation from the APPRISE Washington State
study are quite close. The only FPl rate
available for the County is at the 100% level.
Were the rates available for the 125% and 150%
levels, the derived households total would
certainly be higher.
In a look at the sub-ounty level, Figures 5.2,
5.3, and 5.4 show each tracts at-risk
households as a share of total households in
2008, using the 125/150% adjusted FPL The
tract numbers are from the 200 census.
Appendix A provides the definition of the
principle cit/town abbreviations shown in
parenthesis for each tract.13 The tract are
arrnged frm highest to lowest shares, with
the overall county share of 24% ARH shown as a
red line in each graph.
The City of Spokane, reflecng Its size, contins
the largest number of tract; however It also
contains the tract with the largest shares of
ARH. Most of these tract are locted In the
central, east-eentral, and northeast portions of
the city. The Cit of Cheney (Figure 5.4) also
has a relatively high percentage of ARH. In the
case of both east-eentral Spokane and Cheney,
this may reflec, In part, the influence of the
universit populations associated with Gonzaga
and Eastern Washington Universities.
Figure 5.2: At-risk Households in City of Spokne Ara at the 125/150% FPL Adjusment in 200
9l
80
70
ll
i 50i-
~ 30
j 20.
10%
, ~ofAl~~Il~%,
200 Ce.. Tra(AClawn)
16
Ex No.7ca No. AVl-10-1 and AVl10-1D. Ko, AviSCul 2, Pag 22 of 81
Figure 5.3: At-rk Housholds in Ci of Spone Valley Area at the 125/150% FPL Adjusment In 20
40%
35%
30%
15 25%
S
i 20%
~ 15%
j 10%
co
5%
0%
~i ~~!!!!!!.......N ....N........
'~êiAf..c:":M""
i ~f L I f â f i i f â
!!!!!!
~
!!!!
I0...==~..N N ..NN..q q q q q :i........z .!!.::.N
I
N N N N l....l'........l'..
!!!!
==
N
..i2Oc..... Tr(ACit_n)..
Figure 5.4: At Risk Households In Other Count Ars at the 12/15 FPL Adjusment in 20
10%
50%
I... ..~~..Co~N%1
15 40
S.s 30
I
'I 20%
jco 10%
0%
f¡f¡l íi g î ~I íi I g i'íi i I l i f¡II I z
lzZII!!~l 0 ~II Z
S!S!a ~l a S!..N ..q .=l'...I i:....
I
...qqq!q N ....II l'=
..q N l'
~~l'0 II ....0 l'l'....0 ..a II 0 ..i.......~..q ..!!!.....:i:..
2Oc... Tr(A. DC"" Cit_n)
Figure 5.5 (Graphs 5.3 and 5.4) shows the
absolute number of at-risk households with a
householder age 65 or over (ARH65) in 1999,
2003, and 200, as well as their share of total
65 and over households (HH65), at the
125/150% FPL adjustment and at 60% of
household median income adjustment. As of
2008, there were approximately 9,40 ARH65 at
17
Exib No.7
Ca No. AVU-E-1Ð-1 an AVl1Ð-1D. Ko, Avi
Scul 2, Pa 23 0181
the 125/150 FPl adjustment and 17,00 at the
60 adjustment. In share terms, these
estimates reflect 26% and 55% of total HH65.
Alhough the estimated number ARH65s has
increased, the share of ARH65s to total HH65s
has not changed significantly since 1999.
Figure 5.5: Estmates of the Number of At-rik Housholders 65 and Over in 1999, 200, and 200
Gr 5.3: Tot AR
2500
5,00
-1.-28
-208
2000
1,5,00
¡
l! 10,00
125150 X 1.5 Pen FPL 80 X WA II lI
Gni 5.: AR si. ofTot HH
10%
-,.
-203-20
60
50
18~40
l'õ 30
jlØ
120
0%
125150 X 1.5 Pen FPL 80 X WA Med Hl
18
Exbi No.7
case No. AW-E-10.1 an AW-G10.1D. Ko, AviSC 2, Pa 24 of 61
5.3 The Correlation beeen AfH and ARH65
Provides a Reasnable Esmate of the
numbers of AfH65
Since one of L1HEA's target demographic
groups are at-nsk households with individuals
65 years and over, it useful to explore the
correlation between the share of all at-risk
households (ARH) and at-risk households with a
member age 65 or over (ARH65) by census
tract. Figure 5.6 is a scatter graph of ARH and
ARH65 for all trëlct in Spokane County in 1999.
Figure 5.6 clearly shows a positive and
significant correlation between ARH and ARH65.
The correlation coeffcient between ARH and
ARH65 is 0.74. In other words, tract with a
high share of ARH also tend to have a high share
of ARH65. This implies that if the share of ARH
is the only available indicator for an area (e.g.,
tract, counties, cities, or states), then the share
of ARH can also be used as an indicator of ARH
5.4 Th Share of SNA UHEA and URA
Households to esimated AR
Figure 5.7 (Graphs 5.5 and 5.6) shows th
total number and share of household receiving
heating assistance through L1HEAP and L1RAP
since 200, as distributed by SNAP" (Recll that
heating seasons are labeled by the year in
which they end.) SNAP households (SNAP HHs)
receivng energ assistnce through
these programs are also compared to the
estimated at-risk households and total county
households (Graph 5.7).
Figure 5.6: 1999 Relatonship beeen At-rk Househlds & At-nsk Householders 65 and Ovr at th
125/150 Adjustment
lI
lI
I 70
:z
J 60
S 50
18i 40
is
;30.ifn
I 20..
10%
0%
0%
..
.
.
.
.
.
. ....
70 lI lI10%20 30 40 50 60
199 s.. of AR to Tot lØ
19
Exbi No.7
Ca No. Avu10-1 an AW-G10-1D. Kop, Avi
SC 2, Pa 25 of 61
As mentioned in section 3.1, SNAP currently
uses a 125% adjustment for establishing
household eligibilty for heating assistance.
Graph 5.5 records a dramatic increase in SNAP
households, one that reverses a downward
trend that started in 2006. Note, however, that
the number of lIRAP recipients increased only
slightly in 2009. Nearly all the increase in the
past heating year was attributable to L1HEAP.
The path of the two most recent heating years
depicted in Graph 5.6 reveals that L1HEAP and
L1RAP accounted for about 70% and 30% of
SNAP HHs in 2008, respectively. Due, however,
to an increase of funds through the 2008/200
federal stimulus program, L1HEAP's share
increased to 80%. Graph 5.7 shows that SNAP
served about 22% of at-risk households at the
125%/150% adjustment and 14% at the 60%
adjustment in 2008.
Over the years covered, SNAP has covere 22-
26% of the eligible households, as measured by
the FPL rule. It has covered 14-16% of eligible
households, as measured by the median
household income rule.
Graph 5.7 shows that relative to all COunty
households, SNAP-assisted households have
accounted for about 5% of households. If the
200 projection of ARH (discussed In the
following section) Is used as a base, then SNAP
may have served around 30% of eligible
households at the 125%/150% adjustment and
19% of eligible households at the 60
adjustment.
20
Exbi No.7
ca No. Aw--1o-1 an AVU-G1o-1D. Ko, Avi
Scul 2, Pa 26 of 61
Figure 5.7: SNA UHEAURA Houhold frm 200.200
Gni 5.: Tot SN HH by Pn
14.GO
1z.~ 13,1.
l 10,GO
11.l5
.7 10,10,3~--9.8,183 ~8,~---'7.t:I 8,GO:I 8,~IL 8,i 5,80 --5,8,GO
I ..-3,4.181 4.GO
4,GO ...w --2,62 2,..1..-
2,GO 20 20 20 20 20 20v-
I ..NA LI .._SNA LI .._SNAURA..I
Gni 5. Sli of SN HH by Pn
80%SØ
80%
70%
7~
~-12 12 58 ~
r 80IL.t 50 41%:I ....3S 38%:I 4O
~-""29
i 3O ~2O'I 2O
j 10%
0%20 20 20 20 20 20
Vn,
I _SN LI HH.._SN UR HH I
Grn 5.7: Sl of SN HH (LIEALRtoAR ii Tot HH
30
:I 25:I
1 2O ....
:I
~
'I 15%!....i 10%:I:I
~5%zfI
0%20 20 20 20 20v-
I
-s HH S"- of AR 12!150 X 2.5 P_ FPL -SNA.. S.. ofAR IO XWA.. ..
- SNA HH S.. of TDl ..
21
Exibi No.7ca No. Aw--10-1 an A~10-1D. Kop, Avi
SCedle 2, Pag 27 of 61
The 2005 UHEAP evaluation study found about
13% of eligible households in 2001 received
lIHEAP assistance. SNAP's higher share of
households served reflect the decision to set
the local eligibilty threshold at 125% of the
applicable FPl, which is below lIHEAP's
. legislated maximums, and Avista's lIRA
dollars.
5.5 Projecion of ARH for 200-2012
Section 5 is concluded by considering annual
projections of Spokane Countys at-risk
households (ARH) for the 200-2012 periods.
Two different approaches are used to generate
projections for each of the three povert line
adjustments. The first approach generates
projections as follows:
(9) Ro= ((H08)(l + gt)(w08,a)
for t = 1,...,4 at povrt adjustment a = 125/150;
and 60 of HHI.
Where:
Ro is the projected number of the Countys ARH at
time 2008+t;
H08 is the total number of OFM ocupied households
in 200;
g is the average annual geometric growt rate of
occupied households from 1999-200 (g=O.012); and
W08a is the estimated share of ARH in 200 to total
HH in 2008 at povert level adjustment a.
This "fixed share method" assumes that W08,a is
a reasonable approximation for Wøa given
that t is not large.
The second method uses regression analyis to
map the relationship beteen 1999 median HHI
and the 1999 share of ARHs (W99) in each of
Washington's 39 counties, again using the
125%/150% and 200 adjustments. The
regression equation is used to estimate W for
each county is as follows:
(10) W.J,99,a = bo + bi(HHIJ,99) + b2(HHIJ,)2
for j = 1,...,39 counties at povert adjustment a=
125/150% and 60 of HHI.
Where:
W. j.,a is the regression estmate of Wj. at povert
adjustment a;
HHIj. is county j's median household income in
1999.
These estimated regression equations are
shown in Figure 5.8. Each of these two
regression equations is then used to estimate
the countys. future share (W.) by simply
projecting forward HHI deflated to 199 dollars.
In this case, the HHI projection for this forecast
is generated by taking the average of real
median HHI (in 1999 dollars) over the 199-
2007 period for Spokane County (the average
used is $37,90).XYThis average of HHI
projecion for 2008 to 2012 is inserted into
equation (8) to generate W.jt for 2008 to 2012.
22
Exbi No.7
Ca No. AW-E-1Ð-1 an Avu1Ð-1D. Ko, AviSC 2, Pag 28 of 61
Fleure 5.8: Recreion Relationship beeen ARH and Median Household Incoe acros 39
Washlnetn Contes In 199
Blue · 125150 X 2.5 FPL Or. 80 X WA llan lH
80
50
:i:i)40 .I!
S
130
'õ
120.cfI
10%
W= 2E-10(HHI)2 -3E-05(HHI) + 1.1439
R2= 0.9906
W" 3E-1O(HHI)2 -3E-G5(HHI) + 1.319
R2= 0.93
0%
25 27,00 21.0 31,00 3300 35 37.0 38 41,00 43 45 47.0 48.0 51.0 5300 ss
199 Median HH Ince
Therefore, equation (9) becomes:functional form expressed by equation (10) has
not fundamentally changed since 1999.
(11) Rot= ((H08)(1 + g)t)(W.¡.t,)
for t = 0,...,4 at poert adjustment a = 125/150
and 60 of HHI,
Where t = 0 is the projection for 200, t = 1 for 200,
and soon.
Table 5.1 presents the projections from both
methods. Both methods generate similar
results, and suggest that at the 125/150 FPL
adjustment and the 60 of HHI adjustment, the
number of ARH wil Increase by approximately
550 and 850 households per year, respeively.
The larger household growth associated with
the 60 adjustment reflect the impact of the
larger eligible base estimated for 200.
The advantage of equation (11) is that it can be
used to simulate the impact of changes in real
household income (HHI) on a countys ARH.
However, this approach assumes that the
23
Exibi No.7
ca No. AYUE.1Ð-1 an AYU1Ð-1D. Kop, Avi
SCul 2, Pa 29 of 61
Table 5.1: Projecons of At Risk Households in 5pokane Count for 2002012
Year
200
200
2010
2011
2012
43,016
43,533
44,055
44,584
45,119
69,706
70,542
71,389
72,246
73,113
\~-ø~,~~¡:'
516
522
529
535
836
847
857
867
200 44,532 69,299
200 45,066 70,130 534 832
2010 45,607 70,972 541 842
2011 46,154 71,824 547 852
2012 46,708 72,685 554 862
24
Exibi No.7Ca No. Avu.1D-1 an Aw-1D-1D. Kop, Avi
Sc 2. Pag 30 of 61
6. Analysis of Spokane County
Households Assisted by LlHEAP &
LlRAP
SNAP's database of lIHEAP and lIRAP
recipients in Spokane County provides
both a cross-sectional and time series
picture of the heat burden borne by low-income
households. Each observation in the database
represents an individual household and can be
broken out by household characteristics, such
as the presence of children 0-5 years (HH5);
adults 60 years or older (HH60); handicapped
individuals (HHHC); and household location by
zip code. In addition, as the elecronic database
goes back to 200, the time dimension of heat
burden can be examined.
However, it is important to remember that the
database does not consist of a single household
cohort followed each year. The SNAP records
contain those households who qualifed for
lIHEAP assistance, and they mayor may not be
in multiple years of the database. Also, since
income is self-reported on a monthly basis, a
reporting bias of an unknown size is likely
reflectd in the data. To convert monthly
income into an annual estimate, each
household's reported income is multiplied by
12.
Since all SNAP recipients show high heating
expenditures relative to their income, the rates
in this section are all expressed as burdens. As
noted in section 3, analysts of the national
lIHEAP program estimate that a Nhigh" burden
occurs when heating and cooling cost are
greater than or equal to 4.3% of gross HHI,
while a "moderate" burden is more than 2.6%
of HHI, but less than 4.3% HHi.16 Therefore, if
the heat burden of a SNAP-assisted household
is above these thresholds, then it is likely that
their energy burden for heating and cooling is
higher.
6.1 Summary Stastcs of Heat Burden, 20-
200
Tables 6.1 through 6.3 provide summary
statistics for each of the relevant groups of
households in heating years 200, 200, and
200. (In the study, heating seasons are labeled
by the year in which they end.) Following theNindivdual" methodology of the 200 lIHEAP
evaluation study, all burden statistics in Tables
6.1-6.3 are based on calculating heating
burdens for individual households and then
calculating the median burdens for each group
under consideration. With the exception of
household size, the median, rather than mean,
is used because a comparison of the mean and
median of gross HHI, heating costs and heat
burden showed relatively skewed distributions
in all years. As a result, the median is a beer
measure for characterizing a "typical" SNAP
household. Recall that the median reflect the
heat burden that 50% of households are above
and 50% are below.
The two measures of heat burden are grs
heat burden and net heat burden. The gross
heat burden is calculated for only those
households that report positive income, and
reflec the heat burden in the absence of
energ assistance. In contrast, net heat burden
is calculated as annual heating cost less energy
assistance, divided by gross HHI. Finally, the
sub-roups in each table are not mutually
exclusive, in that some households with
children may also be represented in the
households with adults 60 years or over or
handicappe persns.
2S
Exbi No.7
ca No. AW-E-1G-1 an AW-G1G-1D. Ko. Avi
SCul 2, Pa 310161
Table 6.1: Analyis of SNAP (UHEA+URA) Housholds, Heatinl5eason 20
Number of HHs
Share ofHHs
Mean HH Size
L1HEAP HH Median Monthly Income, $
L1HEAP HH Implied Annual Median
8,785
100%
2.72
5,60 1,521 959 2,406
64%61%66%66%
2.65 4.15 1.56 2.02
3,181 991 485 1,215
36%39%34%34%
2.85 4.17 1.71 2.25
580 633 556 550
585 632 583 558
572 633 516 550
790 1,059 721 708
9,482 12,705 8,648 8,492
710 1,067 712 68
9,242 12,800 8,54 8,250
829 1,047 729 743
9,94 12,564 8,748 8,916
467 530 437 435
460 513 452 431
481 538 421 442
Number of L1HEAP HHs
Share of All HHs
Mean L1HEAP HH Size
Number of L1RAP HHs*
Share of All HHs*
Mean L1RAP HH Size*
All HH Median Annual Heating Bil, $
L1HEAP HH Median Annual Heating Bil, $
L1RAP HH Median Annual Heating Bil, $*
All HH Median Monthly Income, $
All HH Implied Annual Median Income, $
L1RAP HH Median Monthly Income, $*
L1RAP HH Implied Annual Median Income,
Median Annual L1HEAP+L1RAP HH Benefit,
Median Annual L1HEAP HH Benefit, $
Median Annual L1RAP HH Benefit, $*
;~':'T~:~¡:~r~f~r i:w~~~~jr;1"~r,7fE~j!Jj ~~:-: :;,,,.~~H ::;~-~~,cl::~7;t::"~\~, ~",~;," '~"/~~, ':~~..~~ :~~ ~ ~: ~ ~T: '" ~ :,S', '# ~~'f f '\,~~"'-'Ji ~: ~t .~, :i~:': )1lt~~:i~.~,.xi¡;~~,g,1~2.,~"",.%V~w..'!~it..~~~..'" ;¡~,-~~~..~P~-"~'i'\. ~",,¥e;,¡;fli1,~ '",Jt_ "h-";¡~ '''£c.. _..~';,::"':::y ..:;~.., ...;,,; ,.~_,~ .. k ii ~.,\ ~$;
All HH Gross Median 5.4%4.7%5.9%5.6%
All HH Net Median 0.8%0.7%1.1%0.9%
L1HEAP HH Gross Median 5.5%4.6%6.2%5.8%
L1HEAP HH Net Median 0.9%0.7%1.1%1.0%
L1RAP HH Gross Median*5.2%4.8%5.6%5.3%
L1RAP HH Net Median*0.8%0.7%1.0%0.9%
* L1RAP households include 216 households that received an Avista energ tax rebate in lieu of a traditional L1RAP
subsidy.
26
Exib No.7ca No. Aw--10-1 an AVl10-1D. Kosk. AvISC 2, Pa 32 of 61
Table 6.2: Analyis of SN (UHEAURA) Households, Heating Seasn 200
~êJJlmr:Z1¿Ü~~:"i!£S~tj~ r~~~:~f~~ t£~~~;;: ;~~~,~:~~~~~~~;~'~~~ ~yi;~:.,~~- ~ ~~; ~:~t ~~ ;¿~ ,~ L~~~~e,~:~:~::: :: ~~' j~ ~~, \
Number of HHs
Share of All HHs
Mean HHSize
Number of L1HEAP HHs 6,569 1,688
Share of All HHs 71%71%
Mean L1HEAP HH Size 2.51 4.18
Number of L1RAP HHs 2,624 676
Share of All HHs 29%29%
Mean L1RAP HH Size 2.54 4.21
All HH Median Annual Heating Bil, $695 783
L1HEAP HH Median Annual Heating Bil, $701 790
L1RAP HH Median Annual Heating Bil, $686 769
All HH Median Monthly Income, $870 1,259
All HH Implied Annual Median Income, $10,44 15,102
L1HEAP HH Median Monthly Income, $872 1,264
L1HEAP HH Implied Annual Median 10,464 15,162
LlRAP HH Median Monthly Income, $860 1,251
L1RAP HH Implied Annual Median 10,324 15,006
Median Annual L1HEAP+L1RAP HH
Median Annual L1HEAP HH Benefit, $
Median Annual L1RAP HH Benefi, $
519
516
529
608
611
604
1,308 3,139
71%71%
1.46 2.00
535 1,308
29%29%
1.51 2.02
684 669
692 682
630 637
789 780
9,468 9,360
786 781
9,437 9,372
796 775
9,552 9,300
469 477
483 474
441 487
All HH Gross Median 5.9%5.0% .6.5%6.1%
All HH Net Median 1.4%1.1%1.8%1.5%
L1HEAP HH Gross Median 6.1%5.0%6.7%6.3%
L1HEAP HH Net Median 1.5%1.1%1.9%1.6%
L1RAP HH Gross Median 5.7%4.9%6.0%5.8%
L1RAP HH Net Median 1.3%1.1%1.5%1.4%
r;~!!~?~~~~~~Kr$'~~~,~~~~-:i~~~_l¡t~r, ~~ t ~~::~~::~,;:'l~~ ,:~,'~,,", ~'~':--t;)-:': ~~~:~~ , ~ ~ ,:', ;'~"" ,:¡ ~:t ,: ~~? ~ : -~~_;:' 1~~~""~.a~1!L.1a¡ti~~~,,t.~. "æ""-5~"o\~W _",",_"J ~.."',.~ ;;VL"""~":f~~J..:~~~ ¡¡",¿~""'l;#.,'1'" l"~t _ J" ~.."~,-"".,,~$' è&J:,~A _"""," ~ l "Atgw;¡1f "~"i~",ik l~
27
Exibi No.7ca No. AVUE-1G-1 an AVl1G-1D. Ko, Avi
SChul 2, Pag 33 of 81
Table 6.3: Analys of SNAP (UHEAP+RA) Househods, Heating Seson 200
Number of HHs
Share of All HHs
Mean HHSize
13,140
100%
2.59
2,326
18%
1.53
2,725 2,067 4,64
75%89%85%
4.19 1.50 1.99
890 259 821
25%11%15%
3.96 1.78 2.27
854 710 715
851 710 710
866 792 750
1,292 803 821
15,500 9,637 9,852
1,291 812 803
15,495 9,744 9,637
1,296 931 919
15,554 11,172 11,033
650 499 507
652 492 500
64 54 54
Number of lIHEAP HHs
Share of All HHs
Mean lIHEAP HH Size
10,459
80%
2.55
Number of lIRAP HHs
Share of All HHs
Mean lIRAP HH Size
2,681
20%
2.75
All HH Median Annual Heating Bil, $
lIHEAP HH Median Annual Heating Bill, $
lIRAP HH Median Annual Heating Bil, $
774
767
794
All HH Median Monthly Income, $
All HH Implied Annual Median Income, $
931
11,172
lIHEAP HH Median Monthly Income, $
lIHEAP HH Implied Annual Median
907
10,88
L1RAP HH Median Monthly Income, $
L1RAP HH Implied Annual Median Income, .
1,030
12,360
Median Annual L1HEAP+L1RAP HH Benefit,
Median Annual L1HEAP HH Benefit, $
Median Annual L1RA HH Benefit, $
557
549
586
l!~tít~:rlt~it~l~~~~'~\e:i:~:~:~~ki¿~~~~:i~:~~~:~~:~¡~i;,i:*i~ J ~~h~~~r;; ~:~~;~ ~\~~:_Z;~ ~~:.~, ~~~~, ~~~~~~
All HH Gross Median 6.1%5.3%6.9%6.5%
All HH Net Median 1.4%1.1%1.9%1.6%
L1HEAP HH Gross Median 6.2%5.3%6.9%6.6%
L1HEAP HH Net Median 1.4%1.1%1.9%1.6%
L1RAP HH Gross Median 5.9%5.4%6.5%6.2%
L1RAP HH Net Median 1.4%1.2%1.8%1.5%
tmr~£f~~~1f~;~r;y~:~~ :_7t~~~~-l\~::~;"'\~~2'~~S¡4L~~::1~"1::;-:-1 - ~~~t"':1:~~g~~.t~~",j1~~ ~ \ ~ ,:~J;F ~fi: l~ ~.l'~~t ~.:~~'4~~~\il;lil:kJ.ék:¡~k~:~~~1~",'£t.:.l;~,,~'t:~~~.t.3t&~-~¡,~~~~...'~~,, '" *i~k"~.J~li~~.~~n.."..~ ~
28
Exibi No.7
ca No. AW-E-1()1 and AVU-G1()1D. Kopsk, Avi
Scle 2, Pa 34 of 61
An inspection of the summary sttistics above
the grey bar in Tables 6.1-6.3 reveals few
differences between the financial
characteristics of SNAP, L1HEAP, and L1RAP
households. This is not surprising, as L1RAP's
eligibilty rules are the same as for L1HEAP. The
similarity between the two participants of the
two programs also extends to each of the sub-
categories.
An examination of the calculations in the lines
below the grey bar of Tables 6.1 and 6.2 shows
an increase in the median gross heat burden
between 200 and 2008. Between the two
heating seasons, the gross median heat burden
for all SNAP-assisted households increased from
5.4% to 5.9%, or about 0.5% points. A similar
increase in the median heat burden was
observed for L1HEAP and L1RAP households, and
for each of the three sub-groups. This is not
surprising given the run-up in energ prices
between 200 and 2008.
A comparison of Tables 6.2 and 6.3 depict only
slight increase in gross heat burden between
heating seasons 2008 and 200. However, since
employment declines accelerated in the spring
and summer of 2009, the recession's impact on
HHI (the denominator of heat burden) may not
be fully captured by the 2009 data. As a
benchmark, the gross median heat burden for
all U.S. L1HEAP households in 2006 was 7.1%
(200 L1HEAP Home Energ Notebook, p. 7,
Table 2-4). This suggests that the heat burden
of SNAP households is lower by about one
percentage point.
Within the sub-groups, HHGO and HHHC have
higher burdens, compared to all SNAP
households and those with very young children
(HH5). Nevertheless, all of the groups show a
gross median heat burden higher than the 4.3%
threshold defined by L1HEAP as a "high" heating
and cooling burden. This suggests that in the
absence of L1HEAP, the typical SNAP household
would be seerely stressed if all energ costs
were considered.
Tables 6.1 and 6.2 also demonstrate that
between heating seasons 200 and 200 the
net median heating burden for all categories
increased approximately 1.6 times. This reflec
an increasing share of heating costs to gros
household income (gross burden), and a stble
or declining share of L1HEA benefit to gross
HHI. Nevertheless, in both years, the median
net burden was less than the lower end L1HEAP
threshold of 2.6% that defines a moderate
burden. In the past two heating seasons, there
was litle change in net heat burdens.
Table 6.4 presents a comparison of four growh
rates for each category of SNAP-assisted
households for the 200 and 200 periods: in
median annual HHI, the growth in median
annual energy assistnce benefi, in the total
growth of the median annual heating bil, and in
the CPI inflation rate for Western blc cities.
Over 200200 (the period of rising energ
prices), the median heating bil for all SNAP
households grew by 20% while median HHI only
grew by 10%. A similar pattern exist for the
three sub-etegories of SNAP-assisted
households.
For all SNAP-assisted households and for each
sub-ategory, the median assistnce benefit
grew more slowly than or just kept pace with
median HHI. Finally, with the exception of
households with young children (HH5), inflation
exceeded median HHI growth, which suggests a
general contraction in the budget constraints of
SNAP recipients. (CPI inflation, excluding
energy cost, grew at or below median
household income growth.) Although some
caution is needed because the SNAP data do
not follow a single cohort through time, the
evidence suggests a material deterioration from
200 to 200 in the financial position of a
typical SNAP recipient household.
29
Exib NO.7
Ca No. AVU-E-10-1 an AVl10-1D. Ko, Avi
SCe 2. Pag 35 of 61
Table 6.4: Groh Analyis of SN (LiHEAP+URAP) Houseolds, Heatins seasns 200 to 200
Annual Median Income 10%
10%Median Annual Energ benef 11%
24%23%22%Median Annual Heating Bil 20%
.ca~d
'.5_:.'.':
19%9%10%
15%7%
CPI West b/e Index 13%
CPI West b/e Index, Le Energ 10%
~~:l~~l: :i~~~~~~~;~:'~'~~~:;'~: ':'~"~~~~~7E :~.~::~ 1'¡~~1g;~I#~
~;l~':~;¡l¡,;:~fi.ifff~t:'J~~ ~~..;, ;;":~iv5-~tt, ~~1t~~;r~~¡~%..?i¡¡~%tf;t;1''lJl~-':"r;'jzf~~¡hl'::~ ~ 'l'g:'-'.'7!f'i;";lf')~~t t.,,-:__ ~~. 5-;J~¡""'h~ ~~l' ~~t~hi:tf:t"-i~~~,;1q,';s", ,-"~;jgif!:,,,~~ :';¥""v'f"'''j'.~''''':!:::Wif'' "!f;;"'".\Ìf;\lnr.ltir;:r~Tl ~~!J ~\(("""""l i:"¡ilj ~"". .--.?J .~ , ;&;~~,;:~;a~"'':~~~dí y~ ""~~~ !ó~~~~~dA .'IY" '" . ,) f ~~;;J,\""1
Annual Median Incme
13%
10%
13%
10%
13%
10%
ttWIØtÖte:
S:Yrs,: .
7%3%2%5%
Median Annual Energ beneft 7%7%6%6%
Median Annual Heatins Bill 11% 9% 4% 7%
CPI West b/e Index (1st Half of 20009) -1% -1% -1% -1%
CPI Wes b/e Index Less Energ (1stHalf of 200-() 2% 2% 2% 2%
Note: Growth rates for median income, energ benefrt, and heating bil are calculated by taking the percentage
change from 200 to 2008 and 200 to 200.
An examination of the two most recent heating
seasons (a period of a deepening recession)
reveals that the median heating bil for all SNAP
households grew by 11%, while median
household income (HHI) grew by 7%. As before,
a similar pattern also exists for the three sub-
categories. Unlike the 2002008 period
however, the median assistance benefit grew
faster or just kept pace with median HHI for all
categories, while inflation was below median
HHI growth. In fact, the CPI data for the firs six
months of 200 shows deflation, due in large
part to decline in energy prices. Excluding
energy, consumer inflation is running around
2%, which is at or below median HHI growth.
Whether or not this is providing any real budget
relief to Spokane County at-risk households
depends on the how strongly the
unemployment (or underemployment) impact
of the recession are being felt.
In addition, as was noted above, because the
SNAP data do not follow a single cohort through
time, the robust income growth (7% for all
SNAP households) over heating season 200
may reflec the combined impact of higher-
income households seeking energ assistnce
due to the recession and the recent expansion
of the assistance dollars. In fact, a careful
examination of Tables 6.2 and 6.3 reveals that
30
exibit No.7
Ca No. AW-E-1G-1 and Avu1G-1D. Kop, Avi
Scul 2, Pa 36 of 61
from 200 to 200, the median HHI of all L1RAP
households increased by 20% while median HHI
of all L1HEAP households increased by only 4%.
As a result, the median monthly income
differential between L1HEAP and L1RAP
households went from -$59 in 200 and +$12 in
2008, to -$123 in 2009. That is, L1RAP
households show significantly higher monthly
income in 2009. This means, unlike previous
years, L1RAP dollars in the most recent heating
season were more frequently allocated to
households with incomes higher than those
funded with L1HEAP dollars.
6.2 Distrbutional Anlyis of Heat Burden by
Household in Three Recent Heatlne Years
To obtain a better picture of the range and
distribution of heat burdens, Figures 6.1 and 6.2
examine the distribution of heat burdens across
all SNAP HH by individual households and
geographic location. Here, geographic location
is defined by a SNAP HH's five-digit zip code.
Figure 6.1 (Graphs 6.1 and 6.2) shows the
cumulative frequency distribution for gross and
net heat burdens in 200, 2008, and 200.
Here, a cumulative frequency distribution
shows how quickly the total number of SNAP-
assisted households increases as the heating
burden increases. A flatter slope of the line, as
in Graph 6.1, indicates that it isn't until a gross
burden of 12% that the vast majorit, say 90%,
of the households are accounted for.
Equivalently, the remaining 10% of SNAP-
assisted households reported a gross heating
burden greater than 12% in 200. A steeper
slope to the line, as in Graph 6.2, implies that
the vast majority of SNAP recipients faced a low
net heating burden in all three heating years.
For example, about 90% of SNAP recipients
showed a net heating burden of less than 2% in
200. With an adequately-funded program and
accurate qualification of households, a
diference in slopes of the gross and net heating
burden curves should be the outcome.
Note further that in a cumulative frequency
distribution, 50% on the vertical axis
corresponds to the median heat burden on the
horizontal axis. For example, imagine taking a
pencil and placing its point on 50% on the
vertical axis, and then drawing a horizontal line
straight across to the black line (representing
200). Next, imagine drawing a line straight
down from this point on the black line to the
horizontal axis. On the horizontal axis the
pencil would touch the median heat burden for
200, where 50 of households are above and
below this number (the median value is shown
in Table 6.1). The same process could also be
applied to the orange and blue lines which
reflect heating seasons 200S and 200. Finally,
also note that the last heat burden bin (unit) in
Figure 6.1 is for all burdens more than 50.
Figure 6.1 reveals that both the gross and net
heat burdens significantly shifted to the right
between the 200 and 200 heating seasons. In
other words, the burdens increasd for SNAP
recipients. Beeen the last tw seasons, there
was slight rightward shift in the gross burden,
while the net heat burden was litle changed.
Between 200 and 200, most of the shif in
gross heat burden occurred in the 4% to 25%
burden range; for net heat burden, the range
was 1% to 15%. Boh of these shift are
consistent with the median changes in Tables
6.1-6.3, and imply higher heat burdens were felt
by more than 90% of the households. As of the
most recent season, about 69% of SNAP
households had a gross heat burden in exces of
4.3% while 10% had a net heat burden in excess
of 4.3%. In 200, these same values where 61%
and 6%, respectively.
31
Exbi No.7ca No. Avu-1D-1 an Avu1D-1D. Ko. AviSc 2, Pa 37 of 61
Figure 6.1: Cumulatie Fruency Distibution of Heang Burden of SNAP Housholds in Heaing
seasns 200, 200 & 200
Gr 6.1: Gr Hu Bu, 8I. 20 Or · 20 MI Blue. 20
10%
/~U
t,
jJ
100
90
80
f 70
l 80
i 50¡'S 40
E
d 30
20
0%0% " ~ ß ß ~~~~~20~~2020~~~~~40~~~~80
Gr He Burd
Gr'6. NI Hu Bu 8I. 20 Or · 20 MI Blue. 20
100
90 ~..rr80Ill70l.
r fl
u.i 50l
E 40
E
d 30
20
10%
0%
0%2% 5% 7% .%11% 14% 1ß 11% 20 23% 25 27 21 3" 34Y, 36% ~ 41Y, 43% 45Y, 47% 80
NI Hu Burd
Note: 200 is the 2003-0 heating season, 200 is the 200-08 heating season, and 200 is the 200-0 heating
season.
Figures 6.2 and 6.3 show a distributional
analysis by zip code. Figure 6.2 shows the share
of SNAP households in each reported zip code
in the same three heating seasons, starting
32
Ex No.7Ca No. Avu-1l)1 an AVt1l)1D. Kop. Avi
SCle 2, Pag 38 of 61
from the zip code with the largest share of
households in 200. Figure 6.3 shows the
median heat burden in each reportd zip code
in the three seasons (Graph 6.3), starting with
the zip code with the largest share of
households in 200 (see Figure 6.2). Graph 6.4
reproduces Graph 6.3 to zip code 9931
(Spangle, WA area). Appendix A provides the
definition of city/town abbreviations shown in
parenthesis for each zip code.17
Graph 6.2 reveals that in all three seasons, the
top four and top nine zip codes account for
approximately 50% and 75% of SNAP-assisted
households, respectively, and are largely
located in the City of Spokane. It also shows
there has been little change in the zip code
shares between 200 and 2009. The top 15 zip
codes represent the urban core areas of the Cit
of Spokane and Spokane Valley. The remaining
codes reflect the less urbanized areas of the
County.
Figure 6.3 show that Spokane County zip codes
with the highest share of SNAP households als
have the lowest heat burdens. That is, starting
from the first zip code (99207) there is a slight
upward trend in the median heat burden in
both years. This suggest tht household heat
burden is slightly higher in less urbanized areas,
perhaps reflecting diferences in housing and
heating options, as well as income earning
opportunities.
Some caution is needed in interpreting the
median heat burden in zip codes after 9931,
however, since the number of assisted
households in each of these zip codes is very
small-typically five or fewer households.
Nevertheless, Graph 6.4 clearly shows this trend
out to zip code 9931. This suggests that rural
and urban households may face different heat
burdens and, therefore, urban household heat
burdens cannot necessarily be use to directly
infer the level of rural heat burdens.
Figure 6.2: . Distibution of SNAP UHEAP+RAP) Hosehods by Zip Co in Heating seasns 200-20
25
20
:i:i
~15'
Z(I
'õ!10%..c(I
5'
0%
!
I
..
j
I t1
r ~-~..a
-
....-2l-2l
~~~~~~~~~~~~~~~~~~il-s~~~~-a ~läw~~~w~~I-l ~2~~~~~¡l~~~äLLLLLLLLLLLWLLLLEL L ~ z06~~ ~ ~L QLLL$ L~L
~;~!~~;~~;~l~~;S!i- ;l~a._l ~lisgla~ ~g;Il~~~l ~ee~
111111!il !IIII i (lillKI ai151111111111 ¡ i I ~I Ii III i II; IIIIiii I I
L___ ZI
Cod, Ra by Sha of SNA HH In 20
Note: 200 is the 2003-0 heating season, 2008 is the 200-0 heating season, and 200 is the 200-0 heating
season.
33
Exbi No.7
Ca No. AVU-E-10-1 and AYU10-1D. Ko. Avi
Sc 2. Pag 39 of 61
Fliure 6.3: Median Heat Burden of SNA (LiHEAP+RAP) Households by Zip Code in Heaini seass
200, 20, and 200
c 25 ,Il
i 20
I. 15%
l I
c 10% Ii i5%
Grh 6.: Meia Gr Hea Burdn30
!
0%
--aM----
~~~~~~s~~s~~~~~s~~~-s s~~~i-d ~1~~~~l~~~ll-g2 £~22~-l~~~~~~~~~~~~~~~~W~~~~EL ~ ~ ~~~6 ~ æ~ dL~~ L~Leee8e8~~8~8!88~1~IlaSaw~ _~ ~li~~ia~ g~Il8~8 l888d
i I I í l l Ii I Ii i I l iii it I il it iii i II i l II i IIII i III i I! III i i
ZI Cod Raed by Sh.. of SNA HH In 20
II 14%12%Ie, .
1'2 10%IiI.
I! 8%, .
ie 8%: G
I j 4%
II ¡
I 2% 1
i ft
I
Grph 6.: Meia He Burn to ZI Cod 991
--2O----
~~~~~~S~~S~K~~~~~~~-S ~~2~~-d KI~-!l ~~~LLL~~~~~~W~L~LEL L ~ ~ G888ee8~88~8B888S~Ilssli~E L ~lli I I í II Ii Iii II l i lí¡iiiiii iiliil i
Z1coce, Ranked by Sh.. of SNA HH In 20
Note: 200 is the 2003-0 heating season, 200 is the 2008 heating season, and 200 is the 2009 heating
season.
34
exib No.7
ca No. AW-E-1G-1 an Avu1G-1D. Ko AvI
Scle 2. Pa 40 of 61
7. Measuring Heating Expenditure
Shares for All of Spokane Count
The goal of this secion is to use cost data
from Spokane Countýs electric and
natural gas utilities to arrie at a measure
of heating expenditure shares for the entire
population dwellng in private residences. While
biling data, stripped of all identifers, were
obtained for households, it was impossible to
match income levels to these records. As a
consequence, all analysis was carried out by
census tract. The results, therefore, of this
section represent census tract averages.
Inferring beyond the averages, say to individual
households, is highly problematic. For the
purposes of this section, heating expenditures
shares are calculated as the total average
residential energ expenditures for space
heating over the 2007-08 heating season as a
percentage of the average of 2007 and 2008
median household incomes.
7.1 Metlogy" Data
The analysis generally proceeds by firs
calculating total expenditures on energy for
heating purposes, or a heating surcharge, for
every census tract. The label surcharge is
adopted to indicate household energy used for
space heating only, over all other uses. The
total is expressed as an average heating bil for
all households in the census tract. That result is
then placed over the tracts median household
income to arrive at a ratio that expresses what
the typical household in that census tract might
spend on heat as a share of its income.
7.1.1 Gas and Elecri Heatng Expenditure
Esmations
Natural gas monthly billng information was
proided by Avista Utilties at the census tract
level. Avista shared the total number of natural
gas customers and the total natural
expenditures in each census tract. If a
residence had natural gas service, it is assumed
that the residence uses natural gas as its main
heating fuel. Only natural gas used for heating
purposes was included in the estimations.
To determine this subset of natural gas use
required the identification of a "base month", a
month where virtually no natural gas was used
for heating. An examination of the average
residential gas use led to the choice of June as
the most likely month to have litle energy use
for heating. The sum of biling diferentials for
the months of October through May, versus the
prior June, during the 2007-08 heating season
then constituted the heating surcharge.
Residential electric monthly biling information
was provided by thre of the five utilities that
serve the county: Avista, Oty of Cheney, and
Inland Power and Light. The three gave this
information either by census tract, Zip+4 Code,
or street address. Biling information that was
provided at the Zip+ Code level and street
address was sent to Bamberg-Handley Inc., a
geocding service that assigns the most likely
census tract based on address information. The
three utilities included in the analysis
represented nearl 88 percent of the residential
market share in heating season 2008.
To arrie at an estimate ofthe amount of
electricity spent for heating purposes required
the identifcation of a similar base month, a
month where virtually no electricit is used for
heating or air-conditioning purposes. After an
examination of the average residential elecric
use for the three utilities, June was again found
to be the month with the lowest average total
energy use per residential customer, thus the
most likely month to have little energ use for
heating or cooling. The sum of June billng
differentials for the months of October through
May, versus June, during the 2007-08 heating
season consituted the heating surcharge for
these electricit users.
The average surcharge for households heating
with electric and gas for each census tract was
then calculated by the following method:
35
Exbi No.7
ca No. Avu.1G-1 and AVU-G1G-1D. Kop, Avi
SCul 2, Pa 41 of61
· Multiply the number of households by the
respective average household heating
surcharge for every census tract to arrve at
the total heating surcharge for the
following:
o Households with Avist gas service
o Households with Avista elecricity
service but no Avista gas service
o Households with electricit service
from utilties other than Avista.
· Sum the total heating surcharges calculated
for these three types of residences
· Divide the total heating costs by the sum of
these three types of residences
The heating expenditures of those households
served by utilities that could not provide census
tract-level data were approximated by the
average cost of Inland Power & light
residential customers. The number of these
households was restricted to census tract that
lie in the zip codes served by these utilties.
It should be noted that heating expenditures for
households heating with oil or propane are not
included in the estimates above but the number
of households are. At this point, the calculated
heating surcharges for census tract are
underestimated. The following steps attempt
to estimate oil and liquid propane gas (LPG)
heating use in each tract.
7.1.2 Fuel 011 and Uquld Prpane Gas Heanc
Expenditure Esmatns
No information on the number of oil and LPG
users for the 2007-08 heating season was
available. The research team consequentl used
result from the 200 census and adopted the
simplifing assumption that the numbers had
not changed in the intervening years. As such,
oil and LPG households were assumed to
constitute 7.4% and 1.5%, respectively, of
County households during the 2008 heating
season.
Average heating cost for those residences had
to be calculated, since the research team did
not have access to customer billng data frm
the Countýs providers of these fuels. However,
an estimate could be made from national data.
According to the U.S. Energ Information
Administration (EIA), residences in the Western
U.S. heating with oil or propane spent a total of
$1,592 and $2,04, respectively, during the
period of October 1, 2007 through March 31,
2008; households in the West heating with
natural gas spent an average of $591 during the
same period. Table 7.1 takes up a complete
profile of historical seasonal expenditures by
heating fuel and regin.
Table 7.1 Average Residential Heating Co by Fuel Tye
Our cost estimates assumed that the EIA
Western 07-08 cost ratios of oil and propane to
natural gas applied equally to each census tract
in the County. Spècitically, oil and propane
were calculated to be 2.7 and 3.5 times the
cost, respectively, of natural gas. Average
household natural gas heating expenditures
determined for each census tract were then
multiplied by these ratios to estimate the cost
of households heating with oil and LPG over the
2008 heating season. These calculated average
costs were, in turn, multiplied by the presumed
number of households heating with each of the
two fuels to arrive at total heating expenditures
for the two fuels in each census tract.
7.1.3. Overall Heating Co Esimation and
Heatng Burden by census Tract
The following puts the above steps together.
1) Subtract the total number of households
heating with oil or propane from the
calculated number of households heating
with electricit for each census tract. The
balance is the estimated total number of
households that heat with electricity.
2) Weight the average heating cost for each
fuel by the respective number of
households in each census estimated to use
the fuel for:
· Natural Gas
· Electricit
· Oil
· Propane
3) The result is the total heating surcharge for
that census tract.
4) Divide this result by the number of
households in each tract to arrive at the
Average Heating Surcharge. More formally,
the calculation for the Average Heating
Surcharge (SC) for any census tract j is
calculated as:
SCi = (NpsSCg +Noiu + NLPSCo +
NA~AvltaEl + NRESCø IIN¡
Where:
N1 = Total number of households wiin the census
tract using fuel ii
SCi = The average surcarge witin the census trct
for fuel i.
Where the fuel subscipts are:
gas refers to Avist natural gas
households
oil refers to fuel oil households
LPG refers to liquid propane gas
households
AvElecric refers to households using Avista
eleccity but no natural gas
REA reers to households using
elecricit from non-Avlst utilties
5) To calculate the average heating share of
the tract, divde overall average census
tract heating cost for all fuel types
combined by the average of median
household incomes in 2007 and 2008 for
each census tract.
We note that the number of households using
non-Avist elecricity was adjusted downward
by the number of households using wood as a
heating source. While the number of wood
users has likely retreated since the most recnt
count (census 200), its size, at approximately
5,00, was too big to ignore. As with the fuel oil
and LPG estimates, the 200 number served for
the 200 heating season estimate.
7.2 Result
Appendix B contains the result, tract by tract.
Table 7.2 below summarizes the results by
ranges of average heating shares.xvlI The
heating share values largely correspond to the
"group" mean measures displayed in Table 3.2.
Note that the number of households has ben
reduced by approximately 5,00 from the OFM
estimates for 2007 and 200, since wood-
37
Exbi No.7ca No. Aw--10-1 an Avu10-1D. Ko, Avi
Scul 2, Pag 43 of 61
Table 7.2 A Summary of the Freuenc Heating Share in All census Tract in Spokane Count,
For Heating Season 200
:. 4.0"1 1,630 0.9"
3.0%:.4.0%1 1,04 0.6"
2.0%:.3.0%9 16,017 9.3"
1.0%:.2.0%71 111,68 64.7"
c: 1.0"24 41,995 24.4"
Totals 106 in,366 100.0%
burning households are outside the purview of
our measurement. It is clear that the vast
majority of census tract produce an average
heating share between 1 and 2 percent. In fact,
the weighted average over all census tract is
1.4". Compare this to the result reported in
Table 3.2 for the entire U.S. for all households
for this measure: 1.1". The difference is
undoubtedly due to lower incomes in the
County versus the national average. It is likely
also due to the use of a median instead of a
mean in the denominator of the ratio.
Note that 11 census tract show an average
heating share greater than 2 percent, but only
two show shares higher than 3 percent. These
results do not imply that everyone within a
census tract faces the average depicted in Table
7.2 and Appendix B. There is undoubtedly a
distribution of income in these tract that puts
some households under these share levels.
However, those tract with relatively high
average heating shares likely have a high
number of households clustered around the
mean. section 5 showed that povert is
clustered in certain zip codes in the County.
As Appendix B reveals, a large range of result
stands behind the groupings in Table 7.2. The
lowest heating share was 0.49" while the
highest was 4.13".
Maps provide an intuitive way to express this
range among the census tract. We thank Avista
for their contribution of GIS softare to provide
the following maps. They are presented in pairs.
The first pair shows our calculation of 2008
heating cost by census tract for the County
and of the City of Spokane. The second pair
shows estimated median household income by
census tract for the County and then the City of
Spokane. The final pair shows the calculated
heating shares for the County and the City of
Spokane.
The heating shares, shown in Figure 7.1 and 7.2,
do not reveal a strong pattern by census tract.
Over all census tract, the estimated heating
expenditure average was $639. Average
expenditures for residential heating tended to
increase somewhat as one moves toward the
City of Spokane core. However, the highest
average residential heating costs are not there
but located on the South Hil, Five-mile, and
Dishman-Mica areas. Average expenditures by
households for heating ranged from $682 to
$1,154 for these census tract for the 2008
heating season.
38
Exbi NO.7Ca No. Aw--10-1 an Avt10-1D. Ko, Avi
SCle 2, Pag 44 of 61
The lowest heating expenditures were
estimated to be in the western census tract of
the City of Spokane and in central north part of
Spokane County. These areas may have a larger
percentage of residences that supplement their
heating with wood, or the dwellngs might be
smaller. In the county overall, 88 percent of
residents used electricity or gas as their primary
heating source according to the 200 census.
Expenditures for the heating season ranged
from $415-$478 for these areas. The Cities of
Spokane and Spokane Valley had census tract
within their boundaries showing heating costs
in this low range as well. This might be due to a
higher percentage of residents who were
apartment dwellers.
39
Exib No.7
ca No. Aw--1Q-1 and AVU-G1Q-1D. Ko, AviSC 2, Pa 45 of 61
Filure7.1
Average
Heating Cost
by Census Tract
Spokane County
N
"+.ns !:.~.../" \..
,.~~..#
-i.:"'!'..
i ." .
I
'"~'.'.~."... ..~" '-y.;' .... . . :i. ~~~.:.:';..::. l)',-.- .~.
. ...~.,... .
".,,,;..
_..........-.,- . AwII.1i e-0......0......f$¿j,......_....-__"11-
'./I' /'0._"
40
Ei NO.7
ca No. AYU-1o-1 an AW-G1o-1D. Ko, AviSce 2, Pa 46 of 61
Ficure7.Z
w+.
s
. 1 J S 4 .
Ili li CoD......D......1./:1......_....--_.....
41
exib No.7
ca No. Aw-.1D-1 an AW-G1D-1D. Ko. AvlSC 2, Pa 47 of 61
An examination of comparative income
information by census tract does point to a
definite pattern. Median household incomes
range from $69,387-$94,296 for households in
census tract located on the upper South Hil,
Dishman-Mica, Five-Mile, Mead, Colbert, and
libert Lake areas. They decrease as the
proximity to the City Core increases. In the city
core, estimated 2008 median incomes for
households ranged from $12,066 to $26,505. In
the surrounding rural areas, median household
incomes fell in the $26,506-$69,386 range, with
households in the municipalities of Deer Park,
Cheney, Medical Lake, and Airway Heights
showing incomes toward the lower end of the
range.
The reported low incomes of Chene and some
Spokane core census tract may be due to the
presence of universities. Students are counted
as households by the U.S. Census and
Washington's Ofce of Financial Management.
42
Exbit No.7
ca No. AYU-1Ð-1 and AVl1Ð-1D. Kopsk, Avi
Scle 2, Pag 48 of 61
Figure 73
Median Income
by Census Tract
Spokane County
wi.
s
43
Exib No.7
ca No; Avu-1G-1 and Avu1G-1D. Ko, Avi
SCle 2, Pag 49 of 61
F1cure7.4
w+.
s
. &aua.,.....0"'--'-012-"-0..--_1I-t--~--
44
Exib NO.7ca No. Avu-1()1 an AVl1()1D. Ko Av1SCle 2, Pa 50 of 61
The display of estimated 200 heating
expenditure shares by census tract, shown
below in Figure 7.5 and 7.6, also yielded a
pronounced pattern. As the proximity to the
Spokane Cit core increases, so does the share
of heat in a household's budget. The highest
heating share was, as noted, 4.13%, and is
located in the inner Spokane City core. Six
adjacent census tract showed shares in the 2-4
percent range. Northeast Spokane City also
revealed some high heating shares.
As one moves out into the suburbs, heating
burdens decreased to 1.5% to less than one
percent. This pattern is exhibited by the City of
Spokane Valley as well, although it is not as
distinct. Two census tract in the City of
Spokane Valleýs "inner-eity showed burdens
of 1.5 to 2 percent, again; however, most fell in
the range of 1-1.5%.
One Cheney census trac and the large swath of
the southern County south are the exceptions
to the pattern of lower heating shares, as one
moves frm the center of the City of Spkane.
However, the qulntile ranking (3rd lowest) of the
households In the southern county matches Its
ranking by household income. The one Cheney
census tract with the 2nd (lowest) qulntiJe
ranking in heating shares also matches its
income ranking. In general, the pattern of
median household income shows a highly
(negatively) correlated relationship with the
pattern of heating expenditure shares. This
underscores the findings of section 5, where
household income levels are seen as a proxy for
heating burden.
45
Exib NO.7
ca No. AYU-11J1 an AVl11J1D. Ko. AvI
SCule 2. Pa 51 of61
Flre7.5
Heating Burden i
by Census Tract for ',- ~
Spokane County jç.
w+N E ,., " ~,j: ,¡ ,s
Ir
J
r'
"-:,.~ .._.._.....".:
I ) \~~. l I-- ¡' ,.' ;i..t-;." -"..f'.i( ,/ .../ /....L.,.,"
J
i(;
....
,,- I'.
......_..... ....
l '.
,.""; ''''-..--. i'i.r,
.'
,.'..g
l"." ".",-. .",
'.'\. i',-,'
~..(
."
..
I~..'
..~ J '~;f.J
': L L--,..,1 ;./ f..: ../......." f.... .I /. ). .'.,.....' r.~'. ....... _I" .- - .,~~. .F .
~" .
........ .ß"".;....
.\\
,-'
i.~"
..J
'. -.. .1
.-.......--/
46
Exib No.7
ca No. AW-E-1o-1 and Avu1o-1D. Ko, Avl
SC 2. Pa 52 of 61
Fllure7.6
\.
Heating Burden
by Census Tract for the Cit of Spokane
-:.... --.....--,
..~..
w+.
s
. , a a .
rr. r . BwDoriO.ri.,..r'~J.,....a_.a.4'_.4'"-.,
//.. -........__11
47
-~
'..
.
~:,.
I ",
i
"\i.~
Ex No.7ca No. AVl-10-1 and A~10-1D. Ko, AviSc 2, Pa 53 of 61
8. caveats, Qualifications &
Conclusions
The essential methodological challenges to this
study lay in the research team's inabilty to
access individual records that contain both
heating cost and income data. In the absence of
this source information, errors of data accuracy
have certainly been introduced. The creation of
separate data sets for household income levels
and heating costs involved a set of unavoidable
assumptions that all contributed some error to
the final result. To estimate household income,
the team worked with income brackets and not
a full distnbution of actual household incomes.
It had to assume that distribution of income
within a census tract did not change over the
near decde under consideration. With no
specific information about household size by
income brackets, it applied an average across all
brackets.
For the estimate of at-risk households with at
least one senior member, the procedure
assumed that the share of a census tracts
senior population was the same in 200 as in
200. For the estimation of the number of at-
risk households over the 200-2012 period, the
techniques employed assumed that future
household population growth wil follow the
rate of the pnor 10 years, that the share of at-
risk households to total households wil remain
constant, or that the relationship between 1999
median household income and the share of at-
risk households throughout Washington State
wil hold in the future for Spokane County. All
these assumptions are subject to change.
The creation of a heating cost data set for the
heating season 2008 for all Spokane County
residences faced many challenges. These le to
the use of several simplifying assumptions. Firs,
not all electric utilties contributed data to the
project. Consequently, costs for the omitd
households had to be proxied by cost from an
appropriate utilty. Seond, the research team
was skeptical of the accuracy of the translation
of elecnc utilit zipcode data into census tract
for certin certain tract. Third, records for
actual fuel oil and liquid propane heating costs
were completely absent. While the later fuel
plays a minor role throughout the County, fuel
oil use is quite high in many, Spokane Cit
census tracts. The cost to County households
had to be inferred from national Department of
Energ data, and not gathered from the
purveyors, as was the case In natural gas and
electricity.
Third and most importantly, with the exception
of Avista natural gas customers, utilit cost data
that the research team received covered a mix
of households that heated with electnclt, fuel
oil, liquid propane, and for non-Avista elecric
utilty customers, Avisa gas. To arnve at a
mutually (fuel) exclusive set of users, the
research team had to use detailed census tract
data from 200 and thereby assumed that the
number of fuel oil and propane users in
2007/200 was the same.
48
Exbi NO.7ca No. AYU-1G-1 an AVU1G-1D. Ko, Avi
SCul 2, Pa 54 of 61
Finally, it bears noting that the resulting heating
share or burden ratio is a hybrid of the group
approach discusse in section 3.3. Its
numerator is a mean, or average, while its
denominator is a median. The measures from
the national survey data reported in Table 3.2
used a ratio of two means. We did not have the
capabilty to calculate median heating cost by
census tract. Census tract household income, as
estimated by the Washington State Ofce of
Financial Management, is published only as a
median. If one assumes a certin homogeneity
within census tract, the difference between
mean and median income, by tract, may not be
great. Normally however, mean income is
higher than median income. If that relationship
holds even slightly within the census tract of
Spokane, then the resulting mean heating
shares or burdens contain an unknown amount
of upward bias.
Despite thes reservations, the research team
notes the relatively high comparabilty between
our results and those from the latest national
survey (RECS). The diferences beeen the two
studies likely rest in the the greater
pervasiveness of povert in Spokane County
than in local data deficits. In sum, the
techniques employed In this study can be
replicated for those service areas in which
annual census tract estimates of popuatlon are
available and in which the natural gas and
electric utilities can provide biling data with
some geocoding.
49
Exib No.7
Ca No. AVUE-1()1 an AVl1()1D. KO, Avi
SCle 2, Pag 55 of 61
References
Avista Utilties, "Low-Income Rate Assistance Program (L1RAP): 6th Annual Report," Spokane, August 29,
2007, submitted to the Washington State Utilities and Transportation Commission.
Applied Public Policy Research Institute for Study & Evaluation (APPRISE), LlHEAP Energy Burden
Evaluation Study, prepared for the Division of Energy Assistance, Ofce of Community services,
Administration for Children and Familes, U.S. Department of Health and Human Services, Princeton,
NJ., July, 2005.
Applied Public Policy Research Institute for Study & Evaluation (APPRISE), Washington State Energy
Needs Final Report, prepared for the Washington Offce of Communit Trade and Economic
Development, Princeton, NJ., December, 2007.
Fisher, Sheehan & Colton, Public Finance and General Economics, On the Brink: 2007, The Home Energy
Affordability Gap, downloaded from http:Uww.homeenergyaffordabiltyap.com/08AboutFSC2.html.
March, 2008.
U.S. Department of Health and Human Services, Administration for Children and Familes, Offce of
Community Services, Division of Energ Assistnce, LlHEAP Home Energy Notebok for Fiscal Year 200,
Washington, D.C., August, 2008.
50
Exibit No.7ca No. AYU-10-1 and AYU10-1D. Ko, Avi
Sced 2, Pa 56 of 61
Appendi A: Key to City and Town Abbreviations in Figures 5.2 and 6.2
AIRW = Airway Heights
CHAT = Chattaroy
CHEY = Cheney
CLAY = Clayton
COLB = Colbert
DEER = Deer Park
EDW= Edwall
ELK = Elk
FAIR = Fairfeld
FORD = Ford
GRNA = Green Acres
lIBLK = Uberty Lake
MEAD = Mead
MICA = Mica
MILL = Milwood
MEDLK = Medical Lake
MRSH = Marshall
MTSPK = Mt. Spokane
NEWLK = Newman Lake
NEWP = Newport
NINE = Nine Mile Falls (Stevens County but associated zip coe includes Washington)
OPPO = Opportunity
OTIS = Otis Orchards
REAR = Reardan (Uncoln County but associated zip code includes Washington)
ROSA = Rosalia (Whitman County but associated zip code includes Washington)
ROCK = Rockford
SPGL = Spangle
SPK = City of Spokane
SPV = City of Spokane Valley
TRNW = Trentwood
VERD = Veradale
VFORD = Valley Ford
51
Exbi No.7
Ca No. AYU-1G-1 an AW-G10-1D. Kosk, Avi
SCle 2, Pa 57 of 61
Appendix B: Table of Heating Shares for Spokane County
Census Tracts in Heating Season 200
Heating
Share of
Number of Esmate Median
census OCupied Househoid Averae Houshold
Tract Housing Unit Income Heatng Co Incme
1 322 $32,384 $562.25 1.74%
2 1,762 $33,127 $682.66 2.06%
3 2,001 $36,999 $60.92 1.62%
4 1,672 $29,757 $561.98 1.89
5 1,40 $4,214 $638.44 1.44%
6 1,175 $4,056 $718.87 1.63%
7 2,059 $42,794 $667.44 L56%
8 1,851 $64,282 $756.19 1.18
9 2,349 $50,252 $772.12 1.54%
10 2,311 $43,630 $85.46 1.96%
11 1,38 $4,376 $80.16 1.65%
12 920 $38,310 $745.79 1.95%
13 1,479 $38,137 $633.96 1.66%
14 2,504 $32,782 $653.08 L99
15 2,012 $34,966 $633.19 1.81%
16 1,389 $26,338 $515.47 L96%
17 1,439 $4,806 $626.19 1.40
18 1,226 $34,40 $562.67 1.64%
19 1,491 $36,990 $740.80 2.00
20 1,749 $35,843 $607.40 1.69
21 978 $38,46 $659.31 1.71%
23 1,992 $29,393 $734.74 2.5O
24 1,029 $17,627 $652.65 3.70
25 2,699 $24,693 $655.69 2.66%
26 1,915 $31,509 $479.39 1.52%
28 339 $30,44 $597.20 1.96%
29 1,191 $43,998 $692.20 1.57%
30 857 $34,882 $614.34 1.76%
31 1,951 $37,797 $655.37 1.73%
32 1,489 $26,211 $722.99 2.76%
33 661 $23,045 $547.83 238
35 1,630 $11,99 $494.82 4.13%
36 2,349 $20,442 $4.09 2.00
38 823 $43,515 $717.11 1.65%
39 945 $4,646 $736.02 1.65%
40 2,535 $30,725 $686.52 2.23%
41 1,04 $47,071 $971.00 2.08
42 1,931 $68,950 $1,146.16 1.66
43 1,382 $67,944 $1,034.20 1.52%
44 1,966 $45,637 $84.02 1.8"
52
Exib No.7Ca No. Avu-1o-1 an Avu1o-1D. Ko, Avi
SCul 2, Pa 58 of 61
Heatng
Esmate Share of
Number of Median Median
Cens OCpie Hohold Averae Houshol
Tract Housing Unit Income Heii Co Incoe
45 1,428 $72,798 $1,143.16 1.57%
461 1,854 $4,64 $789.34 1.77
462 1,113 $51,717 $664.09 1.28
47 2,662 $47,670 $597.29 1.25"
48 1,507 $77,572 $588.53 0.76"
49 2,335 $72,930 $702.90 0.9"
50 1,162 $75,371 $629.85 0.84"
101 1,815 $67,641 $633.83 0.94"
10201 961 $47,748 $650.61 1.3"
10202 1,926 $77,907 $803.16 1.03"
1031 1,301 $39,251 $543.91 1.~
10303 870 $65,490 $60.25 0.93"103 1,472 $4,666 $723.12 1.4
1035 1,654 $66,264 $778.64 1.1ß
1041 1,152 $36,399 $479.61 132"
1042 2,022 $64,796 $620.12 0.96"
1051 2,619 $72,412 $632.62 0.87
10503 2,085 $80,238 $774.85 0.97%105 1,261 $62,614 $629.13 1.00
1061 1,325 $68,924 $65.32 G.
1062 2,434 $89,949 $667.25 0.74"
107 1,428 $88,216 $730.14 0.83"
108 920 $33,938 $44.88 131"
109 1,422 $56,580 $777.38 137%
110 1,333 $47,995 $692.05 1.44"
11101 2,40 $31,838 $534.47 1.&1
11102 1,336 $41,197 $4.35 0.99
11201 2,933 $33,329 $41.64 1.45"
1122 1,437 $51,313 $654.90 1.28
113 2,491 $62,94 $637.39 1.01"
114 1,945 $45,947 $574.97 1.25
115 567 $47,278 $602.43 1.27%
116 734 $41,938 $643.08 1.53"
117 3,455 $35,582 $4.80 L14"
118 2,316 $36,301 $387.69 1.07%
119 1,669 $41,707 $432.97 1.04"
120 1,623 $41,571 $523.34 1.26"
121 1,061 $34,559 $636.70 1.84"
12 963 $38,120 $58.89 1.52"
123 2,376 $35,141 $496.07 1.41"
12401 1,629 $65,933 $643.50 0.9n
12402 1,949 $8,522 $647.59 0.77
53
Ei No.7Ca No. Avu-1Ð-1 an Avu1Ð-1D. Ko, AviSd 2. Pa 59 of 61
Esma
Numbe of Median Averaee
OCupie Household Heatng
Housing Unit Income Co
1,323 $34,80 $452.19
1,467 $45,468 $551.07
1,514 $39,06 $526.07
788 $56,183 $647.38
1,563 $58,273 $523.18
1,281 $61,513 $6.91
1,043 $54,080 $507.85
2,487 $58,215 $393.86
2,499 $50,115 $484.53
3,239 $4,967 $592.68
2,464 $52,039 $617.75
2,824 $73,784 $695.64
813 $64,709 $868.09
1,557 $93,704 $857.36
2,245 $73,993 $732.40
1,293 $4,971 $674.97
917 $46,423 $656.63
1,043 $4,66 $333.52
1,839 $52,750 $546.53
1,942 $23,427 $4.06
1,575 $37,175 $442.88
1,508 $55,747 $652.35
937 $60,258 $292.73
1,035 $47,536 $873.63
54
Ex No.7Ca No. AVl-10-1 an AVU10-1D. Ko, AYi
SCul 2. Pa 60 of 61
Endnotes
1 The Human Servs Amndments of 199, Public Law
103-252, Se. 2602(a), as amended, reportd In the UHEAP
Home Energ Notbok for Fiscal Year 200, U.s.
Dertent of Health & Human Servces, Administti
for Children and Familes, Ofce of Communit Servce,
Divsio of Energ Asistnce, August, 200.
2 From http:Uww.liheapwa.orgPage,aspx?nld=5,
downloded Deceber 15, 200,
3 Sources: Census, Populaon Finder:
http:Ufaetnder.census.gov/servlet/SAFFPopulation? sub
menuld=populatlon 0: Washington State Ofic of
Financial Management:
http:Uww,ofm,wa.gov/pop/aprillldefault.asp: and
Spokane Communit Indicators:
ww.communitlndicators.ewu.edu: all downloaded
12.14.200.
4 Avi Utilities, Six Annual Repo (May 200ril
2(07), submit to the Washingto Stote Utlities and
Transpoation Commision, Augus 29, 2007,
5 Th Western Censu region Includes the Roc Mountain
and Pacifc sttes, as well as Alaska and Hawaii, for a total
of 13.
6 Standard deiation is a measure of the dispeion of a
disribution of numbers, or, how far the values fall frm
the mean. Formally, it Is the square roo of the varince of
a distribution. For data that are highly concntraed
around the mean, the standard devatn wil be low; fo a
widel disprsd disributn, the standard deion wil
be high.
7 From the U.S. Census: "Census tract are small, relativly
permanent sttistcal subdivsions of a county. Census
trct bondaries normally follow visible features, but may
follo governmental unit boundaries and other non-visible
features In some Insnce; they always ne wiin
counties. Designed to be relatly homogneous unit
wih respe to population charact, economic
status, and living conditns, census tra avrage about
4,00 Inhabitnts
8 See the "Conclusions" of the Execuiv Summary.
9 This approach assumes that the share of the 65+
population group In any year sinc 200 has been
relatively constant. While the share has eded up over
time, the movement has ben slight.
10 Fo exmple, asume we have a tol of 16 Incme
bracket per census trct at time t: (Bii.' Bu.i, (Øi,
Bu.i, (Bi3,t, Bu,i, and so on until (Bi.6,t, Bu,i&, If the
adjusd port line fell in bra thre, then the ARH
would be the sum tract houshods In brck on, tw,
and thre.
II 20 UHEAP Energ Burden Evaluati Stdy.
U Spkane Community Indictors,
ww.communitindicators.ewu.edu/graph.dm?id=97
13 Because some tract coer a large area, th ciies/towns
attched to each tract refec th principle poulation
centers In or on th border that tract
14 SNs disribution of funds strt in the fourt quarter
each year and exends into the firs quartr of the Ne
Year. Tht is, strictly speakin& 200 re the winter
months of 2003-0, 2005 reec the winter mohs of
2005, and so on.
xv Th averae was use sinc real annual HHI did no hav
a clear trend over the 1992007 peri. OFM nominal
Income esimates for Spkane County wee us to
calculate this averge, aftr they were defted using th
Wesern CPI for blc cies. Th Index was rele to 50
that th CPI was 100 In 199.
16 Se L1HEAP Energ Burdn Evaluation Study (Final
Report, July 2005), pp. 11-12. This pres a detile
desription of th methodology for calcung th
thresho for high and moderae hetingcoling
burdens.
17 Th ciltow attched wit each zip coe re the
address lottion pro in the SN databa.
ss
Exibit No.7
Case No. Avu-10-1 an AVl10-10, Kopki, Avi
Scle 2, Pa 61 of 61